Viewership Prediction Modeling With Panel-Merged Device Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Content providers face challenges in accurately predicting viewership fluctuations for targeted advertising, leading to inconsistent campaign results due to limited device resources and restricted access to data, resulting in inefficient deployment strategies.
Innovation Solution
A system and method for predicting viewership probability using aggregated data from multiple devices, incorporating machine learning models to generate accurate viewership predictions by merging device data with trusted panel information and demographic data, and comparing model performance against a baseline.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If content providers use broad approach in deploying advertising campaigns across all devices, then advertising reach is maximized, but campaign effectiveness becomes inconsistent due to viewership fluctuations
Solution Approach 1:
The patent segments the advertising campaign deployment by dividing devices into different risk groups (high-risk, medium-risk, low-risk) based on predicted viewership probability. Instead of treating all devices uniformly, the system creates targeted segments that allow consistent campaign effectiveness within each segment while maintaining overall broad reach across multiple segments.
Solution Approach 2:
The system changes the parameter of campaign deployment from a static broad-approach to a dynamic risk-based approach. By using machine learning models to predict viewership probability and assigning risk scores, the campaign parameters (targeting criteria, bid amounts, placement decisions) are adjusted based on predicted performance, transforming a one-size-fits-all strategy into a parameterized multi-tiered strategy.
2Measurement precision
If content providers collect and process data from individual devices only, then data privacy is maintained, but prediction accuracy is insufficient due to limited device resources
Solution Approach 1:
The patent merges data from multiple sources including individual device data, panel data from partnering devices, and demographic data into a unified training dataset. This combination allows the machine learning models to achieve high prediction accuracy by leveraging diverse data sources while distributing the processing complexity across multiple data contributors rather than requiring any single device to handle all processing.
Solution Approach 2:
The system introduces an intermediary data processing layer that aggregates and processes data from multiple devices before feeding it to the machine learning models. This intermediary layer handles the complex data processing requirements centrally, allowing individual devices to maintain simplicity and privacy while still contributing to accurate predictions through their data contributions.
3Productivity
If content providers deploy advertising campaigns without accurate viewership prediction, then campaign deployment is simple and fast, but resource allocation becomes inefficient
Solution Approach 1:
The system performs preliminary action by predicting viewership probability and calculating risk scores before deploying advertising campaigns. The machine learning models analyze historical and current data to forecast which devices are likely to view content, allowing content providers to pre-plan and optimize campaign deployment. This preliminary prediction prevents wasted resources on devices unlikely to view the content while ensuring adequate coverage of high-probability devices.
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring actual viewership outcomes and using this information to retrain and improve the machine learning models. The feedback loop allows the system to learn from past campaign performance, adjusting future predictions and resource allocation decisions. This continuous improvement maintains high deployment efficiency while progressively reducing resource waste as the models become more accurate.
Data Source
AI summary
Approaches provide for predictive viewership associated with a device. Information associated with viewership by the device may be received over an interval. The received viewership information is merged with panel information to further generate merged information. The merged information is then aggregated at a predetermined increment to form aggregated date. The aggregated data can then be used as input training data to a model to generate probability of viewership by the device. One or more metrics associated with the predicted viewership can be tracked to evaluate model performance.