Individual Viewership Attribution via Demographic Probability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video distribution systems fail to identify individual viewers within households, leading to incomplete data for advertising rate setting, as they only report aggregated viewing activities at the device or household level, lacking insight into specific viewer preferences.
Innovation Solution
A system and method that attribute household viewership information to individuals by using viewership data and demographic information to estimate which individuals viewed specific content, incorporating adjustment factors for content popularity and viewing dependencies within households.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If viewership information is aggregated at the household or device level, then data collection is simplified and implementation is easier, but the precision of individual viewer identification is lost
Solution Approach 1:
The patent segments aggregated household viewership data into individual viewer attributions by dividing the household into multiple viewer segments. Each viewer is assigned probability scores for different viewing scenarios based on demographic characteristics, content preferences, and historical viewing patterns. This segmentation transforms undifferentiated household-level data into differentiated individual-level insights without requiring separate measurement systems for each viewer.
Solution Approach 2:
The patent introduces probabilistic modeling and demographic analysis as intermediary processes between aggregated viewership data and individual viewer identification. These intermediaries process the aggregated data through statistical frameworks that incorporate household composition, viewer demographics, and content characteristics to infer individual viewing behavior. This intermediary layer enables individual-level inference from aggregate data without direct individual measurement.
2Measurement precision
If individual viewer identification is implemented directly, then measurement precision is improved, but device complexity and implementation difficulty increase
Solution Approach 1:
The patent creates a universal attribution framework that processes multiple types of input data (household composition, demographic information, content metadata, viewing timestamps) through a single probabilistic modeling system. This multi-functional approach handles diverse data sources and viewer scenarios using consistent mathematical frameworks, reducing the need for separate specialized systems for different viewing situations while maintaining individual-level precision.
Solution Approach 2:
The patent transforms the attribution problem from a deterministic identification task to a probabilistic parameter estimation problem. By changing from binary viewer/non-viewer classification to continuous probability scoring based on multiple parameters (demographics, content preferences, temporal patterns), the system achieves nuanced individual identification while using standard statistical computing tools rather than complex specialized hardware or algorithms.
Data Source
AI summary
A system and method for the assignment of person-level viewership. The system receives viewership information describing the viewing of video content at a household. The system additionally receives demographic information for that household, including the numbers of persons associated with the household. For each combination of viewers, the system calculates the probability that the viewers viewed the content based on the demographic attributes of those viewers and the probabilities that individuals sharing those attributes would view the content. The system then attributes the viewing information to one or more persons from the household based on the calculated probabilities. The system additionally updates the probabilities that individuals having different demographic attributes would view the content based on the selection of persons.