US 10,096,040 B2Granted

Management of the display of online ad content consistent with one or more performance objectives for a webpage and/or website

AssigneeWalmart Apollo, LLC

InventorsShankara B. Subramanya, Girish Natarajan, Guoxun Yuan, Junhao Jiang, Madhusudan Mathihalli, Chris Bolte

Life of this patent

Estimating the term…

Loading patent…

Abstract

Systems and methods are disclosed for managing the display of online ad content consistent with one or more e-commerce objectives. A collection module may be operable to collect attribute values for a set of attributes characterizing user visits to a set of training webpages and subsequent attribute values for a subsequent user visit to a subsequent webpage. A model-generation module may be operable to train a probability model with the attribute values that predicts outcomes for at least one performance metric. A display module may be operable to determine whether to display ad content on the subsequent webpage for the subsequent user visit depending on whether a predicted outcome from the probability model that is relevant to the subsequent attribute values is consistent with one or more e-commerce objectives. The probability model may be a decision tree with different predicted outcomes for different combinations of attribute values.

Claims24

Dependent claims

A system for managing online ads comprising:

one or more processing apparatuses;
and
one or more non-transitory medium storing computing instructions configured to run on the one or more processing apparatuses and perform acts of:
collecting training attribute values, for a set of attributes, from user visits of a plurality of users to a set of online training webpages by recording (1) interactions between the plurality of users and the set of online training webpages, (2) one or more webpages of the set of online training webpages accessed by the plurality of users, (3) time spent by the plurality of users on the one or more webpages of the set of online training webpages, and (4) queries entered by the plurality of users at the one or more webpages of the set of online training webpages;
generating a probability model with the training attribute values providing predicted outcomes for at least one e-commerce attribute in the set of attributes for different combinations of attribute values;
identifying, from the probability model, a first predicted outcome corresponding to a subsequent combination of attribute values collected by a collection module for a subsequent online user visit to a subsequent webpage, the predicted outcomes comprising the first predicted outcome;
coordinating a display of the subsequent webpage comprising either:
(a) an ad at a first location on the subsequent webpage, and a webpage content in a first format at a second location on the subsequent webpage;
or (b) the ad at a third location on the subsequent webpage, and the webpage content in a second format at a fourth location on the subsequent webpage, during the subsequent online user visit where the first predicted outcome satisfies an objective of the subsequent webpage;
and
coordinating a display of an ad-free version of the subsequent webpage during the subsequent online user visit where the first predicted outcome does not satisfy the objective of the subsequent webpage.

8 more elements

A method for ad-display management on a webpage comprising:

collecting a first set of recorded values of at least one performance metric from clickstream data for a first set of online user visits to a set of online training webpages by recording (1) interactions between a plurality of users making the first set of online user visits and the set of online training webpages, (2) one or more webpages of the set of online training webpages accessed by the plurality of users, (3) time spent by the plurality of users on the one or more webpages of the set of online training webpages, and (4) queries entered by the plurality of users at the set of online training webpages, the set of online training webpages configured consistent with an ad configuration during the first set of online user visits;
generating a probability model, with the first set of recorded values, of at least one predicted result for at least one performance metric for future online user visits to the set of online training webpages configured consistent with the ad configuration;
applying a standard to the probability model for an additional online user visit to the webpage;
coordinating a display of the webpage comprising either:
(a) an ad at a first location on the webpage and a webpage content in a first format at a second location on the webpage;
or (b) the ad at a third location on the webpage and the webpage content in a second format at a fourth location on the webpage, wherein (a) and (b) are consistent with the ad configuration where the probability model indicates a relevant predicted result for the additional online user visit to the webpage that satisfies the standard or an objective of the webpage;
and
coordinating a display of the webpage consistent with an ad-free version where the relevant predicted result does not satisfy the standard and the objective of the webpage.

5 more elements

A system for displaying on-line ads, comprising:

one or more processing apparatuses;
and
one or more non-transitory medium storing computing instructions configured to run on the one or more processing apparatuses and perform acts of:
separating user training visits to a set of webpages into a first set of online user visits, the set of webpages displayed without ad content during the first set of online user visits, and a second set of online user visits, the set of webpages displaying the ad content at a first location and webpage content in a first format at a second location, or displaying the ad content at a third location and the webpage content in a second format at a fourth location, during the second set of online user visits;
collecting a first training data set for the first set of online user visits, a second training data set for the second set of online user visits, and a subject data set for an additional online user visit to a subject webpage by recording (1) interactions between a plurality of users making the user training visits and the set of webpages, (2) one or more webpages of the set of webpages accessed by the plurality of users, (3) time spent by the plurality of users on the one or more webpages of the set of webpages, and (4) queries entered by the plurality of users at the one or more webpages of the set of webpages, the first training data set, the second training data set, and the subject data set containing data for a set of attributes that characterize the first and second set of online user visits correlated to user visits;
engaging in decision tree learning on the first training data set and the second training data set to train a decision tree with nodes defined by different values for attributes in the set of attributes and leaves predicting results for a change, in at least one response variable, between versions of the set of webpages displayed without the ad content, displayed with the ad content in the first location and the webpage content in the first format at the second location, and displayed with the ad content in the third location and the webpage content in the second format at the fourth location;
identifying a first predicted result of the leaves predicting the results for the change in the at least one response variable, the first predicted result identified by following the decision tree in accordance with the subject data set;
coordinating a display of the subject webpage with the ad content in the first location and the webpage content in the first format at the second location, or displaying the ad content at the third location and the webpage content in the second format at the fourth location during the additional online user visit where the first predicted result for the change satisfies a business objective of the subject webpage;
and
coordinating a display of the subject webpage without the ad content during the additional online user visit where the first predicted result for the change does not satisfy the business objective of the subject webpage.

7 more elements

Dependencies are inferred from the claim text.