Heterogenous Treatment Effect Model for Persuadable Customer Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Advertisers face challenges in identifying the best source audience for creating effective lookalike audiences, leading to ineffective and costly advertising due to the inability to accurately target persuadable customers.
Innovation Solution
A system and process that utilizes a Heterogenous Treatment Effect (HTE) model trained on survey data to select and generate an optimized customer list with personally identifiable information, identifying persuadable customers for targeted advertising.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional demographic-based audience targeting is used, then advertisers can reach broad audience segments, but the precision of identifying truly persuadable customers deteriorates
Solution Approach 1:
The patent replaces traditional mechanical/demographic-based audience selection methods with a machine learning model (HETE model) that uses survey data and causal inference to predict customer persuadability. This substitution enables precise identification of persuadable customers without relying on simple demographic segmentation, resolving the contradiction between precision and complexity.
Solution Approach 2:
The patent introduces survey data and a HETE model as intermediaries between advertisers and target audiences. These intermediaries enable precise measurement of customer persuadability by mediating the selection process through causal inference, allowing advertisers to identify truly persuadable customers rather than relying on proxy demographic characteristics.
2Quantity of substance
If advertisers use broad source audiences for lookalike modeling, then potential reach increases, but the effectiveness of advertising decreases
Solution Approach 1:
The patent changes the key parameter for audience selection from demographic characteristics to predicted persuadability scores generated by the HETE model. This parameter change allows advertisers to select audiences based on actual likelihood of response rather than demographic proxies, maintaining advertising effectiveness while enabling flexible audience sizing.
Solution Approach 2:
The patent performs preliminary survey data collection and model training before the actual advertising campaign. This preliminary action creates a pre-computed HETE model that can quickly identify persuadable customers in the target audience, ensuring advertising effectiveness is maintained even as audience size varies.
3Measurement precision
If advertisers invest in extensive survey data collection and HETE model training, then identification of persuadable customers improves, but time and resource consumption increases
Solution Approach 1:
The patent performs survey data collection and HETE model training as preliminary actions before the advertising campaign begins. Once the model is trained, it can rapidly score and identify persuadable customers in the target audience without requiring additional time during the actual campaign execution.
Solution Approach 2:
The patent creates a self-service system where the HETE model automatically scores customers and identifies persuadable individuals without requiring manual analysis. This automation reduces the ongoing time and resource investment needed after the initial model training, allowing the system to efficiently process and identify target customers independently.
Data Source
AI summary
A system and process to create a lookalike model for a target audience to deliver advertisements are disclosed. According to one embodiment, the method comprises selecting survey data from a survey database that relates to an advertisement. A heterogenous treatment effect (HETE) model is trained on the survey data. Persuadable customers are identified from the survey database for the advertisement based on the HETE model. An optimized customer list is generated using personally identifiable information.