Demographic Adjustment Modeling for Accurate Audience Impression Records
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing online audience measurement processes face inaccuracies due to self-reported demographic information provided by users during account registration, which lacks veracity and reliability, leading to biased demographic-exposure results.
Innovation Solution
An audience measurement entity collects highly reliable demographic data through in-person and telephonic interviews and installs online meters to monitor panelist interactions, linking this data with self-reported data to model relationships and adjust demographic information using classification tree-based approaches, segmenting data based on behavioral variables to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If self-reported demographic information is collected during account registration, then data collection ease is improved, but measurement precision deteriorates
Solution Approach 1:
The patent introduces an intermediary adjustment model that mediates between self-reported demographic data and actual demographic characteristics. The model uses behavioral variables as intermediaries to infer and adjust demographic information, resolving the contradiction by not relying directly on self-reported data while maintaining ease of data collection through automated behavioral tracking.
Solution Approach 2:
The patent replaces the mechanical system of direct self-reporting with an automated computational system that uses classification tree models and behavioral analysis. This substitution eliminates the need for users to manually provide demographic information while achieving higher measurement precision through objective behavioral indicators.
2Device complexity
If self-reported demographic data is used directly, then device complexity is reduced, but reliability deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-training adjustment models using panel data with known demographic characteristics before deploying them for adjusting self-reported data. This preliminary training phase creates reliable adjustment factors that can be applied automatically, maintaining low operational complexity while ensuring high reliability through pre-validated models.
Solution Approach 2:
The patent implements feedback mechanisms where adjustment models continuously refine demographic estimates by comparing adjusted data against known panel demographics. This feedback loop ensures reliability by constantly validating and adjusting the adjustment factors based on actual performance, while the automated nature maintains system simplicity.
Data Source
AI summary
A disclosed example includes accessing computer-generated impression records, the computer-generated impression records based on network communications received at a server of a database proprietor from client devices, the computer-generated impression records indicative of accesses to media at the client devices; accessing self-reported demographic data and behavioral data from a database of the database proprietor, the self-reported demographic data and the behavioral data corresponding to user accounts registered with the database proprietor and associated with the client devices; comparing the self-reported demographic data with a probability distribution of higher-accuracy demographic data; determining different adjustments for corresponding ones of self-reported demographics of the self-reported demographic data based on the comparison; adjusting the corresponding ones of the self-reported demographics based on corresponding ones of the different adjustments to generate adjusted user demographic data; and assigning the adjusted user demographic data to corresponding ones of the computer-generated impression records.