Predictive Ad Targeting Using Time Series Spending Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional online advertisement systems waste resources by advertising to users year-round, even after they have recently transacted, leading to frustration and unsubscribes, as they fail to predict optimal advertisement windows based on user spending habits.
Innovation Solution
A system and method that uses a computing system to identify customer spending patterns through time series algorithms and machine learning, predicting an anticipated purchase window for users, allowing targeted advertisements to be directed during specific times based on trends, seasonality, and noise in spending habits.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If advertisements are sent year-round to all users, then advertisement coverage is maximized, but resource waste increases and user frustration grows
Solution Approach 1:
The system performs preliminary analysis of user spending patterns and transaction histories to predict optimal advertisement timing windows before sending advertisements. By anticipating when users are most likely to make purchases and respond positively, the system sends advertisements only during these predicted windows rather than continuously, thereby reducing resource waste while maintaining effective coverage.
Solution Approach 2:
Instead of continuous year-round advertising, the system implements periodic advertisement campaigns timed to coincide with predicted user purchase windows. These periodic actions are based on analyzed spending patterns, seasonality, and transaction frequencies, allowing the system to concentrate advertising resources during high-probability periods rather than distributing them uniformly throughout the year.
2Duration of action of stationary object
If advertisements are sent frequently to maintain brand awareness, then marketing visibility is improved, but user engagement decreases due to frustration
Solution Approach 1:
The system dynamically adjusts advertisement timing and frequency based on real-time analysis of user behavior patterns, transaction histories, and predicted purchase windows. Rather than using a static year-round advertising schedule, the system adapts its marketing visibility duration to match individual user spending cycles and preferences, thereby maintaining engagement while preserving visibility.
3Measurement precision
If time series algorithms analyze detailed spending patterns, then prediction accuracy improves, but computational complexity increases
Solution Approach 1:
The system segments the analysis process into distinct components: identifying baseline spending patterns, detecting seasonality effects, and measuring noise levels. By breaking down the complex time series analysis into these separate analytical stages, the system achieves high prediction accuracy while managing computational complexity through modular processing of different pattern types.
Solution Approach 2:
The system transforms raw transaction data into meaningful parameters such as baseline spending rates, seasonal coefficients, and noise thresholds. By changing the parameters from raw data to aggregated statistical measures, the system maintains high prediction accuracy while reducing the computational burden of analyzing individual transaction details.
Data Source
AI summary
A system and method of generating directed advertisements is disclosed herein. A computing system identifies a baseline history of customer spending for one or more merchant category codes. The computing system generates a time series algorithm for each merchant category code of the one or more merchant category codes. The time series algorithm is based on the baseline history. The computing system receives a stream of transactions associated with a user. The computing system inputs the stream of user data into the time series algorithm. The computing system receives, as output, an anticipated purchase window for the user with respect to each merchant category code of the one or more merchant category codes. The anticipated purchase window includes an anticipated date of next purchase and anticipated purchase amount. The computing system directs advertisements to the user in accordance with each respective anticipated purchase window.