Dynamic Model Threshold Adjustment for Audience Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional server log-based methods for monitoring Internet media usage are susceptible to over-counting and under-counting errors due to tampering and caching issues, limiting their accuracy in tracking user access to online media.
Innovation Solution
Implementing a system where media is tagged with monitoring instructions that execute impression requests, allowing client devices to report access data to a neutral third-party audience measurement entity, and leveraging partnered database proprietors to collect demographic data through redirection and cookie access, while adjusting model threshold scores to maintain target user volumes in data segments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If server log-based methods are used to monitor Internet media usage, then implementation is simple and cost-effective, but measurement accuracy deteriorates due to tampering and caching issues causing over-counting and under-counting errors
Solution Approach 1:
The patent introduces a neutral third-party audience measurement entity as an intermediary between content servers and advertisers. This mediator collects media access data directly from client devices through embedded monitoring instructions, eliminating reliance on potentially tampered server logs while maintaining implementation feasibility through automated data collection
2Device complexity
If server logs are used for tracking media access, then no additional client-side infrastructure is needed, but reliability deteriorates due to susceptibility to tampering and inability to detect cached media access
Solution Approach 1:
The patent embeds monitoring instructions within media content before distribution to content servers. These pre-placed instructions are executed automatically when media is accessed from any location (server or cache), ensuring reliable tracking of all media views without requiring complex client-side infrastructure or post-deployment modifications
3Reliability
If model threshold scores are kept fixed to ensure data quality, then measurement reliability is maintained, but adaptability to market demands deteriorates, limiting ability to adjust data segment value
Solution Approach 1:
The patent implements dynamic model threshold adjustment where thresholds are automatically modified based on target user volume requirements for different data segments. This allows the system to adapt to varying market demands and segment priorities while maintaining data quality through controlled adjustment mechanisms that preserve measurement reliability
4Quantity of substance
If partnered database proprietors are accessed to collect demographic data through redirection, then data completeness improves, but device complexity increases due to additional coordination and cookie access mechanisms
Solution Approach 1:
The patent leverages existing database proprietor infrastructure and cookie mechanisms to serve multiple functions: user identification, demographic data collection, and cross-site tracking. By utilizing the universal cookie system already present in web browsers, the patent achieves comprehensive demographic data collection without building redundant identification infrastructure
Data Source
AI summary
Methods and apparatus for adjusting model threshold levels are disclosed. An example apparatus includes memory; and processor circuitry to execute machine readable instructions to: generate an adjusted count of users by applying an adjustment factor based on volume estimates and an updated total volume target to an absolute count of users; and adjust the model threshold based on the adjusted count of users by: determining a decreased model threshold, the decreased model threshold being less than the model threshold; determining a decreased model total based on lookback model score data, the decreased model total corresponding to a number of users that studies the decreased model threshold; and when a first difference between (A) the decreased model total and (B) adjusted absolute count of users is smaller than a second difference, replacing the model threshold with the decreased model threshold.