Machine Learning Target Audience Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for generating target segment audiences based on recency and frequency of website visits are inaccurate due to the use of arbitrary parameters, such as a frequency condition of one and no recency limit, leading to unsatisfactory digital marketing results.
Innovation Solution
A machine learning system is employed to generate recency and frequency parameters by analyzing hit recency and frequency data, allowing for the identification of users who satisfy specific metrics within a target segment, thereby improving the accuracy of target segment audience generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional recency and frequency values are used (frequency=1, no recency limit), then the target segment audience generation is simple, but the accuracy of the target segment audience is poor
Solution Approach 1:
The system performs self-service by automatically generating recency and frequency parameters through machine learning model training. The model trains on historical browsing data to autonomously determine optimal recency and frequency values without manual intervention, thereby improving audience accuracy while managing complexity through automation
Solution Approach 2:
The invention changes the parameters from fixed conventional values (frequency=1, no recency limit) to dynamic values generated by machine learning models. The model produces optimized recency and frequency parameters based on historical data patterns, transforming static parameter selection into adaptive parameter generation that improves measurement precision
2Productivity
If manual experimentation and empirical experience are used to determine recency and frequency values, then the process is simple to implement, but the results are unsatisfactory and time-consuming
Solution Approach 1:
The invention substitutes the mechanical process of manual experimentation with an automated machine learning system. Instead of manually testing different recency and frequency combinations, the system uses computational models trained on historical data to automatically determine optimal parameters, significantly improving productivity while enhancing accuracy through data-driven insights
Solution Approach 2:
The system performs preliminary action by pre-training machine learning models on historical browsing data before actual audience generation. This advance preparation allows the models to quickly generate accurate recency and frequency parameters during execution, improving both productivity and precision without requiring manual experimentation during the audience generation process
Data Source
AI summary
A segment targeting system generates target segment audiences for delivery of digital content. The segment targeting system generates hit recency and frequency summary data, which identifies hit recency and hit frequency for each of multiple cookie identifiers that satisfied a metric in at least one of multiple segments and in a time range in user specified target segment audience criteria. At least a portion of this summary data is used to train a machine learning system. The machine learning system can include multiple machine learning models, with each model being associated with a different false positive to false negative penalty ratio modelling parameter when learning recency and frequency combinations of hits on webpages in each segment that work well to separate cookie identifiers that satisfied a metric for an offering in a target segment and cookie identifiers that satisfied a metric for an offering in other segments.