Ad Content Display Probability Model for Webpage Performance

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Displaying ad content on webpages can negatively impact performance metrics such as dropout and conversion rates, as links in ads direct traffic away from the original website, potentially reducing the webpage's ability to fulfill its objectives, necessitating a balanced approach to determine when ad display is beneficial or harmful.

Innovation Solution

A system and method that use machine learning to generate a probability model based on user visit data, predicting outcomes for performance metrics, and deciding whether to display ad content or an ad-free version of a webpage based on these predictions, optimizing ad placement and format to align with performance objectives.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If ad content is displayed on a webpage, then revenue is generated through remuneration, but dropout rate increases and conversion rate decreases

Engineering Contradiction:
ImproverevenueVSAvoidperformance metrics
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The system dynamically adjusts ad content display decisions based on real-time probability predictions. A probability model evaluates whether displaying ad content will satisfy performance objectives, allowing the system to adapt its behavior (display or not display ads) based on predicted outcomes for each specific webpage and user context, rather than following a static rule

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system uses feedback from the probability model predictions to guide ad display decisions. By continuously evaluating the predicted impact on performance metrics and adjusting display decisions accordingly, the system creates a closed-loop control mechanism that balances revenue generation with performance objective maintenance

Inventive Principle:
Principle #23Feedback

2Quantity of substance

If ad content with links to different websites is displayed, then remuneration is received, but visitors leave the original website increasing dropout rate

Engineering Contradiction:
ImproveremunerationVSAvoidtraffic loss
Core Design Contradiction:
Quantity of substanceVSObject-generated harmful factors

Solution Approach 1:

The system converts the potentially harmful effect of ad links (traffic loss) into a beneficial decision-making process. By using probability models to predict when ad display will or won't cause harmful traffic loss, the system selectively displays ads only in situations where the harm is minimized or avoided, thus converting the harmful potential into a controlled, beneficial outcome

Inventive Principle:
Principle #22Blessing in disguise (Convert harm into benefit)

3Quantity of substance

If ad content is displayed to generate revenue, then conversion rate may decrease, but overall website performance needs to be maintained

Engineering Contradiction:
ImproverevenueVSAvoidconversion rate
Core Design Contradiction:
Quantity of substanceVSProductivity

Solution Approach 1:

The system changes the parameter of ad display decision-making from a binary choice to a probability-based evaluation. By assessing the predicted impact on conversion rate and other performance metrics through probability modeling, the system adjusts its display parameters dynamically, choosing to display ads only when the predicted negative impact on conversion rate is acceptable given the revenue benefit

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS11107118B2Management of the display of online ad content consistent with one or more performance objectives for a webpage and/or website
Publication Date: 2021.08.31 WALMART APOLLO LLC
  • US11107118B2 patent drawing
  • US11107118B2 patent drawing
  • US11107118B2 patent drawing

AI summary

Systems and methods including one or more processing modules and one or more non-transitory storage modules storing computing instructions configured to run on the one or more processing modules and perform acts of collecting training attribute values for a set of attributes; generating a probability model using the training attribute values, the probability model providing predicted outcomes for at least one attribute in the set of attributes; identifying, using the probability model, a first predicted outcome corresponding to a subsequent combination of attribute values collected by a collection module, wherein the predicted outcomes can comprise the first predicted outcome; coordinating a first display of an ad content version of a subsequent webpage when the first predicted outcome satisfies an objective of the subsequent webpage, wherein the first display can comprise: (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; and coordinating a second display of an ad-free version of the subsequent webpage when the first predicted outcome does not satisfy the objective of the subsequent webpage. Other embodiments are disclosed herein.