Using User Context to Select Content

Abstract

In one example embodiment, a method includes receiving a request for content to be presented to a current user. The content has multiple elements that each have multiple possible element variations. The content has multiple possible instances that would result from different combinations of different ones of the possible element variations. The method includes accessing a current-user context of the current user that has data associated with or concerning the current user; predicting multiple responses of the current user to multiples ones of the possible instances of the content; selecting one of the possible instances of the content for the current user based on the predicted responses; and providing for presentation to the current user a current actual instance of the content based on the selected one of the possible instances of the content.

1 . A method comprising, by one or more computer systems:
 receiving a request for content to be presented to a current user, the content having a plurality of elements that each have a plurality of possible element variations, the content having a plurality of possible instances that would result from different combinations of different ones of the possible element variations;   accessing a current-user context of the current user, the current-user context comprising one or more data associated with or concerning the current user;   predicting a plurality of responses of the current user to a plurality of the possible instances of the content based on one or more of:
 the current-user context; 
 the possible element variations of the plurality of the possible instances of the content; 
 one or more previous-user contexts of one or more previous users who have been presented one or more previous actual instances of the content, each of the previous actual instances of the content having been provided for one of the previous users based on one of the possible instances of the content selected for the one of the previous users; or 
 one or more actual responses of the previous users to the previous actual instances of the content; 
   selecting one of the possible instances of the content for the current user based on the predicted responses; and   providing for presentation to the current user a current actual instance of the content based on the selected one of the possible instances of the content.   
     
     
         2 . The method of  claim 1 , wherein:
 the request for content to be presented to the current user was communicated from a mobile device of the current user; and   presentation to the current user of the current actual instance of the content comprises presentation on the mobile device.   
     
     
         3 . The method of  claim 2 , wherein the mobile device is a smartphone. 
     
     
         4 . The method of  claim 1 , wherein the content is a landing page and the elements are landing-page components. 
     
     
         5 . The method of  claim 1 , wherein the content is media comprising audio, video, or both. 
     
     
         6 . The method of  claim 1 , wherein the content is a set of web-search results. 
     
     
         7 . The method of  claim 1 , wherein the content is an advertisement. 
     
     
         8 . The method of  claim 7 , wherein the advertisement is a banner advertisement. 
     
     
         9 . The method of  claim 7 , wherein the advertisement is text-based. 
     
     
         10 . The method of  claim 1 , wherein the content is a set of text-based advertisements. 
     
     
         11 . The method of  claim 1 , wherein selecting one of the possible instances of the content for the current user is further based on one or more business rules pre-determined for the current-user context of the current user, the business rules governing whether particular ones of the possible element variations may be presented to the current user. 
     
     
         12 . The method of  claim 1 , wherein the selected one of the possible instances of the content for the current user is one of the possible instances of the content having a predicted response that satisfies a threshold requirement. 
     
     
         13 . The method of  claim 12 , wherein the selected one of the possible instances of the content for the current user is one of the possible instances of the content having a most desirable predicted response among substantially all the possible instances of the content. 
     
     
         14 . The method of  claim 1 , wherein:
 the selected one of the possible instances of the content for the current user is one of the possible instances of the content used to test a response from the current user; and   the method further comprises:
 receiving an actual response of the current user to the current actual instance of the content provided based on the possible instance of the content used to test the response from the current user; and 
 analyzing the actual response of the current user. 
   
     
     
         15 . The method of  claim 1 , wherein the selected one of the possible instances of the content for the current user is one of the possible instances of the content used as a control instance of the content for analyzing an actual response of the current user to the current actual instance of the content. 
     
     
         16 . The method of  claim 15 , further comprising:
 receiving an actual response of the current user to the current actual instance of the content provided based on the control instance of the content; and   comparing the actual response of the current user to the current actual instance of the content with at least one of the actual responses of at least one of the previous users to at least one of the previous actual instances of the content, the at least one of the previous actual instances of the content having been provided based on at least one of the possible instances of the content selected for the at least one of the previous users so that the at least one of the possible instances of the content each has a predicted response that satisfies a threshold requirement.   
     
     
         17 . The method of  claim 1 , further comprising:
 designating one of the possible instances of the content having a most desirable predicted response among all of the possible instances of the content as a champion instance;   designating one or more of the possible instances of the content that are different from the champion instance as one or more challenger instances, each of the challenger instances having been predicted as possibly being more desirable than the champion instance with respect to the current-user context; and   selecting the one of the possible instances of the content for the current user from the champion instance and the challenger instances.   
     
     
         18 . The method of  claim 1 , wherein the predicted responses of the current user to the possible instances of the content are determined using one or more statistical models that each has one or more inputs corresponding to the current-user context, the possible element variations of the possible instances of the content, the previous-user contexts, the previous actual instances of the content, or the actual responses of the previous users to the previous actual instances of the content. 
     
     
         19 . The method of  claim 18 , wherein each of the statistical models is trained using the previous-user contexts, the previous actual instances of the content, and the actual responses of the previous users to the previous actual instances of the content. 
     
     
         20 . The method of  claim 1 , wherein each of the predicted responses indicates a likelihood of a desired interaction between the current user and the corresponding possible instance of the content or a desired outcome from a possible interaction between the current user and the corresponding possible instance of the content. 
     
     
         21 . The method of  claim 20 , wherein each of the predicted responses indicates a likelihood of the current user purchasing one or more products through the corresponding possible instance of the content. 
     
     
         22 . The method of  claim 1 , wherein:
 the current-user context comprises one or more objective data or one or more behavioral data associated with or concerning the current user;   the objective data comprise one or more of:
 one or more demographic data of the current user; 
 one or more background data of the current user; 
 one or more conditions of a web-browsing session associated with the current user; 
 one or more network-access data of the web-browsing session associated with the current user; or 
 one or more network-traffic data of the current web-browsing session associated with the current user, and 
   the behavioral data comprise one or more of:
 one or more actions taken by the current user; 
 one or more actual responses of the current user to the current actual instances of the content; or 
 a traffic source from which the current user accessed the current actual instances of the content. 
   
     
     
         23 . The method of  claim 1 , wherein:
 each of the previous-user contexts comprises one or more objective data or one or more behavioral data associated with or concerning one of the previous users;   the objective data of each of the previous-user contexts comprise one or more of:
 one or more demographic data of the corresponding previous user; 
 one or more background data of the corresponding previous user; 
 one or more conditions of a web-browsing session associated with the corresponding previous user; 
 one or more network-access data of the web-browsing session associated with the corresponding previous user; or 
 one or more network-traffic data of the web-browsing session associated with the corresponding previous user, and 
   the behavioral data of each of the previous-user contexts comprise one or more of:
 one or more actions taken by the corresponding previous user; 
 one or more actual responses of the corresponding previous user to one of the previous actual instances of the content; or 
 a traffic source from which the corresponding previous user accessed one of the previous actual instances of the content. 
   
     
     
         24 . The method of  claim 1 , further comprising:
 receiving an actual response of the current user to the current actual instance of the content; and   updating the current-user context based on the actual response of the current user to the current actual instance of the content.   
     
     
         25 . One or more computer-readable non-transitory storage media embodying software that is operable when executed to:
 receive a request for content to be presented to a current user, the content having a plurality of elements that each have a plurality of possible element variations, the content having a plurality of possible instances that would result from different combinations of different ones of the possible element variations;   access a current-user context of the current user, the current-user context comprising one or more data associated with or concerning the current user;   predict a plurality of responses of the current user to a plurality of the possible instances of the content based on one or more of:
 the current-user context; 
 the possible element variations of the plurality of the possible instances of the content; 
 one or more previous-user contexts of one or more previous users who have been presented one or more previous actual instances of the content, each of the previous actual instances of the content having been provided for one of the previous users based on one of the possible instances of the content selected for the one of the previous users; or 
 one or more actual responses of the previous users to the previous actual instances of the content; 
   select one of the possible instances of the content for the current user based on the predicted responses; and   provide for presentation to the current user a current actual instance of the content based on the selected one of the possible instances of the content.   
     
     
         26 . The media of  claim 25 , wherein:
 the request for content to be presented to the current user was communicated from a mobile device of the current user; and   presentation to the current user of the current actual instance of the content comprises presentation on the mobile device.   
     
     
         27 . The media of  claim 26 , wherein the mobile device is a smartphone. 
     
     
         28 . The media of  claim 25 , wherein the content is a landing page and the elements are landing-page components. 
     
     
         29 . The media of  claim 25 , wherein the content is media comprising audio, video, or both. 
     
     
         30 . The media of  claim 25 , wherein the content is a set of web-search results. 
     
     
         31 . The media of  claim 25 , wherein the content is an advertisement. 
     
     
         32 . The media of  claim 31 , wherein the advertisement is a banner advertisement. 
     
     
         33 . The media of  claim 31 , wherein the advertisement is text-based. 
     
     
         34 . The media of  claim 25 , wherein the content is a set of text-based advertisements. 
     
     
         35 . The media of  claim 25 , wherein selecting one of the possible instances of the content for the current user is further based on one or more business rules pre-determined for the current-user context of the current user, the business rules governing whether particular ones of the possible element variations may be presented to the current user. 
     
     
         36 . The media of  claim 25 , wherein the selected one of the possible instances of the content for the current user is one of the possible instances of the content having a predicted response that satisfies a threshold requirement. 
     
     
         37 . The media of  claim 25 , wherein the selected one of the possible instances of the content for the current user is one of the possible instances of the content having a most desirable predicted response among substantially all the possible instances of the content. 
     
     
         38 . The media of  claim 25 , wherein:
 the selected one of the possible instances of the content for the current user is one of the possible instances of the content used to test a response from the current user; and   the method further comprises:
 receiving an actual response of the current user to the current actual instance of the content provided based on the possible instance of the content used to test the response from the current user; and 
 analyzing the actual response of the current user. 
   
     
     
         39 . The media of  claim 25 , wherein the selected one of the possible instances of the content for the current user is one of the possible instances of the content used as a control instance of the content for analyzing an actual response of the current user to the current actual instance of the content. 
     
     
         40 . The media of  claim 39 , wherein the software is further operable when executed to:
 receive an actual response of the current user to the current actual instance of the content provided based on the control instance of the content; and   compare the actual response of the current user to the current actual instance of the content with at least one of the actual responses of at least one of the previous users to at least one of the previous actual instances of the content, the at least one of the previous actual instances of the content having been provided based on at least one of the possible instances of the content selected for the at least one of the previous users so that the at least one of the possible instances of the content each has a predicted response that satisfies a threshold requirement.   
     
     
         41 . The media of  claim 25 , wherein the software is further operable when executed to:
 designate one of the possible instances of the content having a most desirable predicted response among all of the possible instances of the content as a champion instance;   designate one or more of the possible instances of the content that are different from the champion instance as one or more challenger instances, each of the challenger instances having been predicted as possibly being more desirable than the champion instance with respect to the current-user context; and   select the one of the possible instances of the content for the current user from the champion instance and the challenger instances.   
     
     
         42 . The media of  claim 25 , wherein the predicted responses of the current user to the possible instances of the content are determined using one or more statistical models that each has one or more inputs corresponding to the current-user context, the possible element variations of the possible instances of the content, the previous-user contexts, the previous actual instances of the content, or the actual responses of the previous users to the previous actual instances of the content. 
     
     
         43 . The media of  claim 42 , wherein each of the statistical models is trained using the previous-user contexts, the previous actual instances of the content, and the actual responses of the previous users to the previous actual instances of the content. 
     
     
         44 . The media of  claim 25 , wherein each of the predicted responses indicates a likelihood of a desired interaction between the current user and the corresponding possible instance of the content or a desired outcome from a possible interaction between the current user and the corresponding possible instance of the content. 
     
     
         45 . The media of  claim 44 , wherein each of the predicted responses indicates a likelihood of the current user purchasing one or more products through the corresponding possible instance of the content. 
     
     
         46 . The media of  claim 25 , wherein:
 the current-user context comprises one or more objective data or one or more behavioral data associated with or concerning the current user;   the objective data comprise one or more of:
 one or more demographic data of the current user; 
 one or more background data of the current user; 
 one or more conditions of a web-browsing session associated with the current user; 
 one or more network-access data of the web-browsing session associated with the current user; or 
 one or more network-traffic data of the current web-browsing session associated with the current user, and 
   the behavioral data comprise one or more of:
 one or more actions taken by the current user; 
 one or more actual responses of the current user to the current actual instances of the content; or 
 a traffic source from which the current user accessed the current actual instances of the content. 
   
     
     
         47 . The media of  claim 25 , wherein:
 each of the previous-user contexts comprises one or more objective data or one or more behavioral data associated with or concerning one of the previous users;   the objective data of each of the previous-user contexts comprise one or more of:
 one or more demographic data of the corresponding previous user; 
 one or more background data of the corresponding previous user; 
 one or more conditions of a web-browsing session associated with the corresponding previous user; 
 one or more network-access data of the web-browsing session associated with the corresponding previous user; or 
 one or more network-traffic data of the web-browsing session associated with the corresponding previous user, and 
   the behavioral data of each of the previous-user contexts comprise one or more of:
 one or more actions taken by the corresponding previous user; 
 one or more actual responses of the corresponding previous user to one of the previous actual instances of the content; or 
 a traffic source from which the corresponding previous user accessed one of the previous actual instances of the content. 
   
     
     
         48 . The media of  claim 25 , wherein the software is further operable when executed to:
 receive an actual response of the current user to the current actual instance of the content; and   update the current-user context based on the actual response of the current user to the current actual instance of the content.   
     
     
         49 . A system comprising:
 means for receiving a request for content to be presented to a current user, the content having a plurality of elements that each have a plurality of possible element variations, the content having a plurality of possible instances that would result from different combinations of different ones of the possible element variations;   means for accessing a current-user context of the current user, the current-user context comprising one or more data associated with or concerning the current user;   means for predicting a plurality of responses of the current user to a plurality of the possible instances of the content based on one or more of:
 the current-user context; 
 the possible element variations of the plurality of the possible instances of the content; 
 one or more previous-user contexts of one or more previous users who have been presented one or more previous actual instances of the content, each of the previous actual instances of the content having been provided for one of the previous users based on one of the possible instances of the content selected for the one of the previous users; or 
 one or more actual responses of the previous users to the previous actual instances of the content; 
   means for selecting one of the possible instances of the content for the current user based on the predicted responses; and   means for providing for presentation to the current user a current actual instance of the content based on the selected one of the possible instances of the content.