Interest-Level Ad Targeting via Trade Zone Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current ad targeting methods, such as behavioral and contextual targeting, are limited in reaching new users and require cookies or specific page keywords, and fail to accurately capture audience interests beyond single-point context.
Innovation Solution
A system that creates 'trade zones' based on offline and geographic data, allowing for interest-level targeting without requiring users to have been seen online or having specific page keywords, by clustering IP addresses and organizing tokens to determine interest levels for targeted advertising.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If behavioral targeting using cookies is used, then ad targeting accuracy is improved, but the system can only reach users who have been seen online before and requires browser cookie participation
Solution Approach 1:
The patent introduces an intermediary mechanism using third-party cookies placed on publisher websites to collect browsing behavior data. This intermediary allows the system to track user behavior across multiple sites without requiring direct browser-based cookies on every site, thereby improving reachability while maintaining targeting accuracy through aggregated behavior data collection
Solution Approach 2:
The patent transitions from first-party browser cookies to third-party cookies as a new dimension of data collection. This dimensional change enables the system to access user behavior data from multiple publisher sites simultaneously, breaking the limitation of requiring users to have been seen online at specific sites before ad requests
2Measurement precision
If contextual targeting using page keywords is used, then ad relevance to current page is improved, but the system is limited to single-page context and cannot capture broader audience interests
Solution Approach 1:
The patent implements preliminary action by collecting and analyzing user browsing behavior data across multiple sites before the ad request is made. The system pre-processes this behavior data to build user profiles and interest segments, which are then used to select ads that match broader audience interests rather than just current page context
Solution Approach 2:
The patent merges multiple data sources including page keywords, user browsing behavior, and third-party cookie data into a unified targeting approach. This combination allows the system to integrate single-page context with broader user interest patterns, enabling ads to be selected based on both immediate page relevance and accumulated user interest data
Data Source
AI summary
A system for improving shape-based targeting by using interest level data is disclosed. According to one embodiment, a computer-implemented method includes creating one or more trade zones, wherein creating a trade zone includes grouping a set of parameters to deliver custom shapes, clustering the custom shapes according to offline data and geographic distribution of IP addresses, and mapping clusters of the custom shapes to IP addresses. Data indicating consumption of a content source is received by one or more trade zones at a calculated rate and the calculated rate is analyzed to determine an interest associated with each trade zone. Targeting is based on a selected trade zone, wherein the selected trade zone is selected based upon a desired interest representative of a desired audience. A targeting request is transmitted including instructions or information associated with a target action, and the target action is performed.