User-based ad ranking

Patent metadata

Field	Value
Publication number	US-9773256-B1
Application number	US-201414477100-A
Country	US
Kind code	B1
Filing date	Sep 4, 2014
Priority date	Apr 18, 2008
Publication date	Sep 26, 2017
Grant date	Sep 26, 2017

Abstract

Advertisement quality measures (e.g., predicted click through rates) are modified in accordance with a user's preferences with respect to domains to which the advertisements relate.

First claim

What is claimed is:

1. A method comprising:
identifying, by a data processing apparatus that includes one or more modules including a learning module, a set of ads for a user session of a first user, each ad having a respective ad quality measure and linking to a respective domain, the identifying including identifying a first ad in the set of ads that links to a first domain;
obtaining, by the one or more modules, data and using the data to perform operations including:
determining, for the first domain, a correction factor specifying a value by which the ad quality measure for the first ad that links to the first domain is to be modified, the value being based on a difference between a non-zero predicted click through rate for ads that link to the first domain when presented during the user session and an actual click through rate for ads that link to the first domain, the predicted click through rate being based on an aggregate performance of the ads that link to the first domain for multiple users;
determining a modified ad ranking score for the first ad based on the ad quality score for the first ad and the correction factor for the first domain;
identifying a different set of ads for a different user session of a second user;
determining that per-user online activity data that has been collected for the second user is statistically unreliable;
in response to determining that per-user online activity data that has been collected for the second user is statistically unreliable, generating a different correction factor using higher level user data, including one or more of a local time of day of a query submission of the second user, a day of week of the query submission, a language of the query submission, a time zone of the query submission, or a geographic location, rather than the per-user online activity data; and
determining a different modified ad ranking score for each ad in the different set of ads using the different correction factor.

2. The method of claim 1 , further comprising:
conducting a selection process using the modified ranking score for the first ad; and
generating and transmitting data that includes instructions that cause the first ad to be presented at a user device associated with the first user.

3. The method of claim 1 , wherein the identifying comprises identifying a second ad that links to a second domain that is different from the first domain, the method further comprising:
determining that the second domain is similar to the first domain; and
in response to determining that the second domain is similar to the first domain, determining a modified ad ranking score for the second ad based on the ad quality score for the second ad and the correction factor for the first domain.

4. The method of claim 3 , further comprising:
determining that the first user has not previously been provided a threshold number of ads that link to the second domain; and
determining to use the correction factor for the first domain to determine the modified ad ranking score for the second ad in response to determining that the first user has not previously been provided the threshold number of ads that link to the second domain.

5. The method of claim 1 , wherein the identifying comprises identifying a second ad that links to a second domain that is different from the first domain, the method further comprising:
determining that there is insufficient information to determine a correction factor for the second domain and the user session; and
in response to determining that there is insufficient information to determine a correction factor for the second domain and the user session:
identifying a third user that is deemed to be similar to the first user and that is associated with a correction factor for the second domain; and
determining a modified ad ranking score for the second ad based on the ad quality score for the second ad and the correction factor for the second domain associated with the third user.

6. The method of claim 1 , wherein the ad quality score for the first ad is based on an aggregate performance of the first ad for multiple users.

7. The method of claim 1 , wherein the identifying comprises identifying a second ad that links to a second domain that is different from the first domain, the method further comprising determining a modified ad ranking score for the second ad based on the ad quality score for the second ad and the correction factor for the first domain.

8. A system, comprising:
a data processing apparatus that includes one or more modules including a learning module; and
a memory apparatus in data communication with the data processing apparatus and storing instructions executable by the data processing apparatus and that upon such execution cause the data processing apparatus to perform operations comprising:
identifying a set of ads for a user session of a first user, each ad having a respective ad quality measure and linking to a respective domain, the identifying including identifying a first ad in the set of ads that links to a first domain;
obtaining, by the one or more modules, data and using the data to perform operations including:
determining, for the first domain, a correction factor specifying a value by which the ad quality measure for the first ad that links to the first domain is to be modified, the value being based on a difference between a non-zero predicted click through rate for ads that link to the first domain when presented during the user session and an actual click through rate for ads that link to the first domain, the predicted click through rate being based on an aggregate performance of the ads that link to the first domain for multiple users;
determining a modified ad ranking score for the first ad based on the ad quality score for the first ad and the correction factor for the first domain;
identifying a different set of ads for a different user session of a second user;
determining that per-user online activity data that has been collected for the second user is statistically unreliable;
in response to determining that per-user online activity data that has been collected for the second user is statistically unreliable, generating a different correction factor using higher level user data, including one or more of a local time of day of a query submission of the second user, a day of week of the query submission, a language of the query submission, a time zone of the query submission, or a geographic location, rather than the per-user online activity data; and
determining a different modified ad ranking score for each ad in the different set of ads using the different correction factor.

9. The system of claim 8 , wherein the operations further comprise:
conducting a selection process using the modified ranking score for the first ad; and
generating and transmitting data that includes instructions that cause the first ad to be presented at a user device associated with the first user.

10. The system of claim 8 , wherein the identifying comprises identifying a second ad that links to a second domain that is different from the first domain, the operations further comprising:
determining that the second domain is similar to the first domain; and
in response to determining that the second domain is similar to the first domain, determining a modified ad ranking score for the second ad based on the ad quality score for the second ad and the correction factor for the first domain.

11. The system of claim 10 , wherein the operations further comprise:
determining that the first user has not previously been provided a threshold number of ads that link to the second domain; and
determining to us

Assignees

• Google Inc

Inventors

• Frumkin Michael
• Gopalratnam Karthik
• Nguyen Khanh B
• Gasser Ralph U
• Friedman Gregory J

Classifications

• G06Q30/0244 Primary

  Optimization · CPC title

• G06Q30/0242

  Determining effectiveness of advertisements · CPC title