Persona Prediction Models Consolidating Multi-Source User Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems lack the ability to effectively monitor and consolidate customer information from multiple sources to predict customer purchases and prioritize high-potential customers, as they fail to aggregate user activities across different identifiers and time windows, leading to incomplete forecasting and marketing strategies.
Innovation Solution
The implementation of machine-learning models that combine user and event data from various sources to identify personas and audiences, using features like event history and user interactions to generate scores indicating purchase probabilities and audience affiliations, enabling better prediction and prioritization of customer purchasing behavior.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current systems monitor customer information from multiple sources, then customer purchase prediction capability is improved, but the ability to consolidate and aggregate user activities across different identifiers remains insufficient
Solution Approach 1:
The patent merges user activities from multiple identifiers (device IDs, cookies, login credentials) into unified persona profiles. The system consolidates disparate data sources including event streams, user profiles, and interaction histories into integrated audience segments, enabling complete view of customer behavior across channels while improving prediction accuracy
2Measurement precision
If user activities are aggregated across multiple identifiers and time windows, then forecasting accuracy is improved, but system complexity increases
Solution Approach 1:
The patent segments the complex data aggregation process into distinct functional components: identifier resolution services that match activities across devices, persona generation modules that create unified profiles, and audience segmentation systems that organize users into targetable groups. This modular segmentation manages system complexity while enabling comprehensive multi-identifier aggregation across time windows
Solution Approach 2:
The system introduces intermediary processing layers including persona identifiers that act as mediators between raw user activities and forecasting models. These intermediaries consolidate complex multi-identifier data into standardized persona profiles, simplifying the input to prediction algorithms while maintaining comprehensive activity aggregation
3Measurement precision
If machine-learning models combine user and event data from various sources, then predictive accuracy is enhanced, but the ability to identify personas and audiences effectively is reduced
Solution Approach 1:
The patent performs preliminary persona identification and audience segmentation before applying machine-learning prediction models. The system pre-processes user data to create standardized persona profiles with consolidated activity histories, then feeds these prepared profiles into ML models. This preliminary action ensures effective persona detection occurs before predictive analysis, resolving the contradiction between model complexity and identification effectiveness
Data Source
AI summary
Methods, systems, and computer programs are presented for estimating a propensity to buy a product or service. One method includes accessing events generated at a website. Each event comprises a data structure describing an operation performed by a user when accessing the website. Further, the method performs operations, for each user from a group of users associated with an audience, comprising: providing event information for a time window, information of the user, and information for a product as input to a propensity machine-learning (ML) model, the model being trained with training data comprising values for features that include event features, user information features, and audience labels; and generating, by the propensity ML model, a score for the user indicating a probability that the user will purchase the product. Further, the method generates a forecast of purchases of the product for the users in the audience based on the scores.