Audience Extension via ML Ranking of Browsing History
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online advertising systems face challenges in effectively expanding audience reach for line items to meet performance and delivery goals, as existing methods rely heavily on manual targeting and lack efficient automation for identifying similar user segments.
Innovation Solution
A system that collects browsing history data, uses machine learning to rank targetable users, and creates new segments by combining ranked users, allowing for automatic audience extension and improved ad delivery by attaching these new segments to line items.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If manual targeting methods are used to identify audience segments for line items, then advertisers can maintain control over targeting criteria, but the system lacks efficiency and automation in expanding audience reach
Solution Approach 1:
The system enables self-service by automatically identifying and creating new audience segments based on behavioral similarities to existing target segments. The machine learning model autonomously analyzes user behavior data, ranks similar users, and generates new segments without requiring manual intervention from advertisers, thus achieving both automation and efficiency in audience expansion.
Solution Approach 2:
The patent replaces manual mechanical targeting processes with an automated machine learning system. Instead of advertisers manually analyzing and identifying similar user segments, the system uses browsing history data and machine learning algorithms to automatically detect behavioral patterns and create new audience segments, substituting human effort with automated computational processes.
2Measurement precision
If advertisers explicitly target users in specific market segments, then targeting precision is maintained, but audience reach is limited and delivery goals may not be met
Solution Approach 1:
The system introduces dynamics by continuously expanding the target audience based on behavioral similarities. Instead of static explicit targeting criteria, the machine learning model dynamically identifies new user segments that exhibit similar browsing behaviors to existing target segments, allowing the audience reach to grow while maintaining targeting precision through behavior-based matching.
Solution Approach 2:
The patent adds another dimension to targeting by incorporating behavioral analysis alongside traditional demographic segmentation. The system analyzes browsing history and user behavior patterns to create new segmentation dimensions, enabling advertisers to reach similar users across different traditional segments while maintaining precision through behavior-based matching criteria.
3Productivity
If new audience segments are created using machine learning models, then audience reach and delivery effectiveness are improved, but system complexity increases
Solution Approach 1:
The system applies segmentation by dividing the user base into distinct audience segments based on behavioral similarities. The machine learning model analyzes browsing history data to identify and create multiple new segments, each representing users with similar behaviors to the original target segment. This segmentation approach improves delivery effectiveness by enabling targeted advertising to each segment while managing complexity through modular segment creation.
Data Source
AI summary
Aspects of the subject disclosure may include, for example, receiving from a campaign manager device information defining a line item in an online advertising system, including receiving information defining constraints for the line item. The subject disclosure may further include collecting browsing history information for targetable users matching the constraints for the line item, generating a machine learning model to rank the targetable users and building a new segment based on users ranked by the model. The subject disclosure may further include providing, to the campaign manager device, a recommendation to add the new segment to the line item, receiving from the campaign manager device an indication to attach the new segment to the line item, and subsequently, providing advertisement content to targeted users according to the line item including the new segment. Other embodiments are disclosed.