ACCOUNTING FOR FEATURES OF PREVIOUSLY-PRESENTED CONTENT ITEMS WHEN SELECTING CONTENT ITEMS FOR AN ONLINE SYSTEM USER
An online system penalizes content items having features matching features of additional content items previously presented to a user within a specified time interval. The online system identifies various features of the content item and identifies features of content items previously presented to the user within the specified time interval. Feature penalties are determined for various features of the content item based on a number of previously presented content items having a common feature with the content item. Weights may be associated with various content items having a feature matching a feature of the content item based on a time between presentation of the previously presented content item and a current time. A penalty for the content item is determined based on the feature penalties for the features of the content item, and the penalty is applied to a bid amount associated with the content item.
1 . A method comprising:
receiving, at an online system, information describing an advertisement request from an advertiser, the advertisement request including an advertisement having one or more features and a bid amount;
identifying an opportunity to present one or more advertisements to a user of the online system;
identifying features of one or more advertisements previously presented to the user of the online system within a specified time interval from a current time;
determining feature penalties associated with the one or more features of the advertisement request, a feature penalty associated with a feature based at least in part on a number of advertisements previously presented to the user within the specified time interval having the feature;
determining a penalty based at least in part on the determined feature penalties;
modifying the bid amount included in the advertisement request based on the penalty; and
including the advertisement request and the modified bid amount in a selection process performed by the online system.
2 . The method of claim 1 , wherein determining feature penalties associated with the one or more features of the advertisement request comprises:
associating weights with each advertisement previously presented to the user having the feature, the weights based at least in part on a time difference between the current time and a time when an advertisement previously presented to the user having the feature was presented to the user;
determining the feature penalty for the feature based at least in part on the weights.
3 . The method of claim 2 , wherein a weight is an inverse function of the time difference between the current time and the time when the advertisement previously presented to the user having the feature was presented to the user.
4 . The method of claim 2 , wherein determining the feature penalty for the feature based at least in part on the weights comprises:
determining the feature penalty for the feature as a sum of the weights.
5 . The method of claim 1 , wherein modifying the bid amount included in the ad request based on the penalty comprises:
reducing the bid amount by a value based at least in part on the penalty.
6 . The method of claim 1 , wherein modifying the bid amount included in the ad request based on the penalty comprises:
determining a ratio of the penalty to the bid amount; and
modifying the bid amount based on the penalty if the ratio is at least a threshold value.
7 . The method of claim 1 , modifying the bid amount included in the ad request based on the penalty comprises
applying a percentage of the penalty to the advertisement, the percentage based at least in part on a ratio of the penalty to the bid amount.
8 . The method of claim 1 , wherein a feature associated with the advertisement request is selected from a group consisting of: an advertising campaign associated with the advertisement request, an object associated with the advertisement request, an advertiser associated with the advertisement request, content included in the advertisement of the advertisement request, a landing page associated with the advertisement of the advertisement request, one or more topics associated with the advertisement of the advertisement request, and any combination thereof.
9 . The method of claim 1 , wherein a feature associated with the advertisement request comprises a type of display unit in which the advertisement of the advertisement request is to be presented.
10 . The method of claim 1 , wherein a feature of an advertisement previously presented to the user of the online system within the specified time interval from the current time is selected from a group consisting of: an advertising campaign associated with the advertisement request, an object associated with the advertisement request, an advertiser associated with the advertisement request, content included in the advertisement of the advertisement request, a landing page associated with the advertisement of the advertisement request, one or more topics associated with the advertisement of the advertisement request, and any combination thereof.
11 . A method comprising:
receiving, at an online system, information describing an advertisement request from an advertiser, the advertisement request including an advertisement having one or more features and a bid amount;
identifying an opportunity to present one or more advertisements to a user of the online system;
identifying features of one or more advertisements previously presented to the user of the online system within a specified time interval from a current time;
determining a penalty for the advertisement request based at least in part on the one or more features of the advertisement request and the features of the one or more advertisements previously presented to the user of the online system within the specified time interval from the current time;
modifying the bid amount included in the advertisement request based on the penalty; and
including the advertisement request and the modified bid amount in a selection process performed by the online system.
12 . The method of claim 11 , wherein determining the penalty for the advertisement request based at least in part on the one or more features of the advertisement request and the features of the one or more advertisements previously presented to the user of the online system within the specified time interval from the current time comprises:
determining feature penalties associated with the one or more features of the advertisement request, a feature penalty associated with a feature based at least in part on a number of advertisements previously presented to the user within the specified time interval having the feature; and
determining the penalty based at least in part on the determined feature penalties.
13 . The method of claim 12 , wherein determining feature penalties associated with the one or more features of the advertisement request comprises:
associating weights with each advertisement previously presented to the user having the feature, the weights based at least in part on a time difference between the current time and a time when an advertisement previously presented to the user having the feature was presented to the user;
determining the feature penalty for the feature based at least in part on the weights.
14 . The method of claim 13 , wherein a weight is an inverse function of the time difference between the current time and the time when the advertisement previously presented to the user having the feature was presented to the user.
15 . The method of claim 13 , wherein determining the feature penalty for the feature based at least in part on the weights comprises:
determining the feature penalty for the feature as a sum of the weights.
16 . The method of claim 11 , wherein modifying the bid amount included in the ad request based on the penalty comprises:
reducing the bid amount by a value based at least in part on the penalty.
17 . A computer program product comprising a computer readable storage medium having instructions encoded thereon that, when executed by a processor, cause the processor to:
receive, at an online system, information describing an advertisement request from an advertiser, the advertisement request including an advertisement having one or more features and a bid amount;
identify an opportunity to present one or more advertisements to a user of the online system;
identify features of one or more advertisements previously presented to the user of the online system within a specified time interval from a current time;
determine a penalty for the advertisement request based at least in part on the one or more features of the advertisement request and the features of the one or more advertisements previously presented to the user of the online system within the specified time interval from the current time;
modify the bid amount included in the advertisement request based on the penalty; and
include the advertisement request and the modified bid amount in a selection process performed by the online system.
18 . The computer program product of claim 17 , wherein determine the penalty for the advertisement request based at least in part on the one or more features of the advertisement request and the features of the one or more advertisements previously presented to the user of the online system within the specified time interval from the current time comprises:
determine feature penalties associated with the one or more features of the advertisement request, a feature penalty associated with a feature based at least in part on a number of advertisements previously presented to the user within the specified time interval having the feature; and
determine the penalty based at least in part on the determined feature penalties.
19 . The computer program product of claim 18 , wherein determine feature penalties associated with the one or more features of the advertisement request comprises:
associate weights with each advertisement previously presented to the user having the feature, the weights based at least in part on a time difference between the current time and a time when an advertisement previously presented to the user having the feature was presented to the user; and
determine the feature penalty for the feature based at least in part on the weights.
20 . The computer program product of claim 17 , wherein a feature associated with the advertisement request is selected from a group consisting of: an advertising campaign associated with the advertisement request, an object associated with the advertisement request, an advertiser associated with the advertisement request, content included in the advertisement of the advertisement request, a landing page associated with the advertisement of the advertisement request, one or more topics associated with the advertisement of the advertisement request, and any combination thereof.