Predicting User Conversion Value via Intent and Value Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Advertisers face challenges in effectively managing marketing strategies and budget allocation due to the complexity of user acquisition, as traditional intent-based systems fail to consider the value analysis, leading to inefficiencies and increased costs in maximizing revenue.
Innovation Solution
Implementing a system that combines intent and value models using machine learning to predict user conversion likelihood and purchase value, leveraging real-time and aggregated data to optimize user acquisition strategies and calculate return on investment (ROI) for marketing campaigns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional intent-based systems are used for advertising management, then advertiser control over marketing targets is maintained, but revenue maximization efficiency deteriorates due to lack of value analysis
Solution Approach 1:
The system segments the advertising management process into distinct functional modules: an intent model for predicting user conversion likelihood, a value model for estimating purchase value, and a bidding system for campaign selection. This segmentation allows each component to specialize in specific tasks, improving overall revenue maximization efficiency while maintaining manageable system complexity through modular architecture.
Solution Approach 2:
The patent introduces machine learning models as intermediary components between traditional advertising systems and revenue optimization goals. The intent model and value model act as mediators that process user data and provide predictive insights, enabling the system to balance advertiser control with automated revenue maximization without requiring complete system restructuring.
2Productivity
If machine learning models are implemented for conversion prediction, then user acquisition efficiency is improved, but computational resources and system complexity increase
Solution Approach 1:
The system performs preliminary actions by pre-training intent and value models on historical data before actual advertising campaigns. This allows the models to be ready for rapid inference during campaign execution, improving user acquisition efficiency while minimizing real-time computational resource consumption. The heavy computational work is done in advance during the training phase.
Solution Approach 2:
The machine learning models are designed to be self-updating and self-optimizing through continuous learning from new data. Once trained, the models autonomously perform predictions without requiring constant retraining or manual intervention, reducing ongoing computational resource consumption while maintaining high user acquisition efficiency.
3Measurement precision
If comprehensive user data analysis is performed, then prediction accuracy is improved, but data processing time and system complexity increase
Solution Approach 1:
The patent applies local quality by focusing data analysis on specific, relevant features rather than processing all available user data uniformly. The intent model and value model are trained to identify and process only the most predictive features for conversion likelihood and purchase value, improving prediction accuracy while reducing data processing time by eliminating irrelevant data processing.
Solution Approach 2:
The system dynamically adjusts data processing parameters based on campaign requirements and data availability. The models can operate with different levels of data granularity and processing depth depending on the specific prediction task, allowing high accuracy when needed while reducing processing time when rapid decisions are required.
Data Source
AI summary
Highly user-specific data is used to calculate user intent to make a purchase and the value of such a purchase. User activity and information is aggregated, per user, for a set window of time and real-time data on recent site behavior is obtained. Aggregated and/or real-time data is considered by a predictive intent model (calculating the probability that the user will make a purchase) and a predictive value model (calculating the expected revenue such a purchase may generate). Weights, specific to each model, are assigned to predictor features tracked in the aggregated and/or real-time user data. The most highly-weighted features of the intent model relate to users' viewing history, and the most highly-weighted features of the value model relate to price and market. By these means, a user conversion value can be obtained, guiding the application of user acquisition strategies for different home sharing markets.