US-12626281-B2 - Systems and methods for adjusting context-dependent parameters in bidding for electronic advertisements

US12626281B2US 12626281 B2US12626281 B2US 12626281B2US-12626281-B2

Abstract

Methods and systems for adjusting context-dependent parameters in bidding for electronic advertisements include identifying a quantity of advertisement displays received by each of a plurality of users during a historical time period, where each user is associated with a random bidding policy. Each user is allocated to a group based upon the advertisement displays received by the user, and a marginal value per dollar spent associated with each different quantity of ad displays received by the plurality of users during the historical time period is estimated. The random bidding policy for each user is adjusted based upon the estimated marginal value per dollar spent. A bid request for an advertisement display opportunity associated with a first user is received, and a bid for the opportunity is determined based upon a characteristic of the opportunity. The determined bid is modified using the adjusted random bidding policy for the first user, and the modified bid is transmitted in response to the bid request.

Inventors

Benjamin Heymann
Rémi Chan-Renous
Alexandre Gilotte

Assignees

CRITEO TECHNOLOGY SAS

Dates

Publication Date: 20260512
Application Date: 20240423

Claims (11)

1 . A computerized method of adjusting context-dependent parameters in bidding for electronic advertisements during loading of a webpage, the method comprising: identifying, by a bid determination computing device, a quantity of advertisement displays received by each of a plurality of users during a historical time period; allocating, by the bid determination computing device, each of the plurality of users to a user group based upon the corresponding quantity of advertisement displays received by the user during the historical time period; determining, by the bid determination computing device, a random bidding factor for each of the plurality of users using a linearized importance weighting estimator algorithm; estimating, by the bid determination computing device, a marginal value per dollar spent associated with each different quantity of advertisement displays received by the plurality of users during the historical time period; adjusting, by the bid determination computing device, the random bidding factor for each user based upon the estimated marginal value per dollar spent for the quantity of advertisement displays received by the users in the corresponding user group during the historical time period, including capping the random bidding factor for each user in a particular user group according to a defined range that minimizes a total cost variation of bids and maximizes a total value of bids; receiving, by an advertisement exchange computing device from a software application of a requesting client device, during loading of a webpage by the client device, a first request for graphical display source code corresponding to a computerized graphical advertisement display to be inserted into one or more advertisement display opportunities on the webpage; transmitting, by the advertisement exchange computing device to the bid determination computing device, during loading of the webpage by the client device, a bid request for an available advertisement display opportunity of the one or more advertisement display opportunities on the webpage, the bid request associated with a first user of the plurality of users; determining, by the bid determination computing device, during loading of the webpage by the client device, a bid for the available advertisement display opportunity based upon at least one characteristic of the available advertisement display opportunity; dynamically modifying, by the bid determination computing device, during loading of the webpage by the client device, the determined bid using the adjusted random bidding factor for the first user; transmitting, by the bid determination computing device, during loading of the webpage by the client device, the modified bid to the advertising exchange computing device in response to the bid request; determining, by the advertisement exchange computing device, during loading of the webpage by the client device, whether to select the available advertising display opportunity based upon the modified bid; and generating, by the advertisement exchange computing device, during loading of the webpage by the client device and based upon the modified bid, graphical display source code corresponding to the computerized graphical advertisement display to be inserted into the available advertisement display opportunity upon completion of loading of the webpage.
2 . The method of claim 1 , wherein the at least one characteristic of the available advertisement display opportunity is a characteristic received from the advertisement exchange computing device that submitted the corresponding bid request.
3 . The method of claim 2 , wherein the at least one characteristic is associated with a publisher of the available advertisement display opportunity.
4 . The method of claim 2 , wherein the at least one characteristic of the available advertisement display opportunity is a URI associated with the available advertisement display opportunity.
5 . The method of claim 2 , wherein the at least one characteristic of the available advertisement display opportunity is a time of day associated with the bid request.
6 . The method of claim 2 , wherein the at least one characteristic of the available advertisement display opportunity is a user identifier associated with the available advertisement display opportunity.
7 . The method of claim 2 , wherein the at least one characteristic of the available advertisement display opportunity is a display format of the available advertisement display opportunity.
8 . The method of claim 1 , wherein the advertisement exchange computing device is a real-time bidding platform, an ad server, a computing device executing an auction for the available advertisement display opportunity within a browser, or a computing device executing an auction for the available advertisement display opportunity within a mobile application.
9 . The method of claim 1 , wherein the marginal value per dollar spent is a marginal cost of ecpm (expected cost per mile).
10 . The method of claim 1 , wherein the advertisement exchange computing device determines that the modified bid wins an auction for the available advertisement display opportunity and generates the graphical display source code corresponding to the computerized graphical advertisement display to be inserted into the available advertisement display opportunity in the webpage upon completion of loading of the webpage.
11 . A system for adjusting context-dependent parameters in bidding for electronic advertisements during loading of a webpage, the system comprising a bid determination computing device having a memory that stores computer-executable instructions and a processor that executes the computer-executable instructions to: identify, by the bid determination computing device, a quantity of advertisement displays received by each of a plurality of users during a historical time period; allocate, by the bid determination computing device, each of the plurality of users to a user group based upon the corresponding quantity of advertisement displays received by the user during the historical time period; determine, by the bid determination computing device, a random bidding factor for each of the plurality of users using a linearized importance weighting estimator algorithm; estimate, by the bid determination computing device, a marginal value per dollar spent associated with each different quantity of advertisement displays received by the plurality of users during the historical time period; adjust, by the bid determination computing device, the random bidding policy for each user based upon the estimated marginal value per dollar spent for the quantity of advertisement displays received by the users in the corresponding user group during the historical time period, including capping the random bidding factor for each user in a particular user group according to a defined range that minimizes a total cost variation of bids and maximizes a total value of bids; receive, by an advertisement exchange computing device from a software application of a requesting client device, during loading of a webpage by the client device, a first request for graphical display source code corresponding to a computerized graphical advertisement display to be inserted into one or more advertisement display opportunities on the webpage; transmit, by the advertisement exchange computing device to the bid determination computing device, during loading of the webpage by the client device, a bid request for an available advertisement display opportunity of the one or more advertisement display opportunities on the webpage, the bid request associated with a first user of the plurality of users; determine, by the bid determination computing device, during loading of the webpage by the client device, a bid for the available advertisement display opportunity based upon at least one characteristic of the available advertisement display opportunity; dynamically modify, by the bid determination computing device, during loading of the webpage by the client device, the determined bid using the adjusted random bidding policy for the first user; transmit, by the bid determination computing device, during loading of the webpage by the client device, the modified bid to the advertising exchange computing device in response to the bid request; determine, by the advertisement exchange computing device, during loading of the webpage by the client device, whether to select the available advertising display opportunity based upon the modified bid; and generate, by the advertisement exchange computing device, during loading of the webpage by the client device and based upon the modified bid, graphical display source code corresponding to the computerized graphical advertisement display to be inserted into the available advertisement display opportunity upon completion of loading of the webpage.

Description

TECHNICAL FIELD The present technology relates to electronic advertisements and, more particularly, to techniques for adjusting context-dependent parameters in bidding for electronic advertisements. BACKGROUND Publisher systems can provide webpages or other online content that can include one or more advertisement display opportunities for computerized graphical advertisement displays (e.g., space for a banner advertisement across the top of the webpage, within an application, or within other media such as videos or images). In some instances, when a user device (e.g., a computer running a web browser) processes a webpage for display, the user device can request, from an ad system, graphical display source code for a computerized graphical advertisement display for one of the advertisement display opportunities. The ad system can provide the graphical display source code to the user device to render and/or display. As part of the advertisement selection process, the ad system communicates with real-time bidding (RTB) computing platforms. The RTB platforms receive bids from various third-party bidding agent systems (also called demand-side platforms (DSP)) that submit bids for the advertisement display opportunity on behalf of advertisers. Typically, the bidding agent systems are responsible for generating a bid for advertisement opportunities that meet the advertiser's requirements, such as cost, value, and audience considerations. The RTB platforms coalesce the bids received from the various DSPs and determine whether the advertisement display opportunity is selected after analyzing the received bids. As can be appreciated, it is largely accepted that showing too many displays to the same user generates ‘display fatigue.’ In other words, the value of one additional display decreases with the number and/or frequency of the previous displays. A common practice in the industry is to use ‘fatigue’ variables in the prediction models, such as counter of past displays on the same user to improve the predictions. However, an optimal bidding policy should also foresee that display fatigue reduces the value of the next displays on the same user-notably because of the users' fatigue. While is true that classical machine learning models can predict how previously won auctions decrease the current opportunity value, these methodologies are not enough to produce a bid that correctly accounts for how winning the current auction shall impact the future values. Indeed, under this perspective, most bidders in existing advertisement bidding systems are impatient, in that they do not fully account for the repeated nature of the auctions. Unsurprisingly, impatience induces a cost—and failure to properly account for this cost results in reduced value and efficiency in generating advertisement displays. SUMMARY Accordingly, there is a need for technology to adjust opportunity value estimates (such as bids) dynamically in the computerized advertisement bidding process to mitigate the cost of impatience using a bidding policy learning step. The methods and systems described herein advantageously exploit data collected with a randomized bidding policy, and modify the policy to account for the future display opportunities of the user. In one aspect, there is a computerized method of adjusting context-dependent parameters in bidding for electronic advertisements. A bid determination computing device identifies a quantity of advertisement displays received by each of a plurality of users during a historical time period, the advertisement displays for each user associated with a random bidding policy. The bid determination computing device allocates each of the plurality of users to a user group based upon the corresponding quantity of advertisement displays received by the user during the historical time period. The bid determination computing device estimates a marginal value per dollar spent associated with each different quantity of ad displays received by the plurality of users during the historical time period. The bid determination computing device adjusts the random bidding policy for each user based upon the estimated marginal value per dollar spent for the quantity of ad displays received by the user during the historical time period. The bid determination computing device receives, from an advertisement exchange computing device, a bid request for an available advertisement display opportunity during rendering of a webpage, the bid request associated with a first user of the plurality of users. The bid determination computing device determines a bid for the available advertisement display opportunity based upon at least one characteristic of the available advertisement display opportunity. The bid determination computing device modifies the determined bid using the adjusted random bidding policy. The bid determination computing device transmits the modified bid to the advertising exchange computing device in response to the