Multi-touch attribution model with energy decay weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current attribution modeling in online advertising inadequately values impressions and other online activities leading up to conversion events, often ignoring preceding events that may contribute to these events.
Innovation Solution
A multi-touch attribution model that calculates the influence of various online activities by analyzing user browsing histories and assigning contribution values to websites based on their impact on conversion events, using an energy decay concept to weight the significance of interactions over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If Last Touch Attribution Model is used to simplify computation and explanation, then ease of operation is improved, but measurement precision deteriorates because most information from advertising campaigns is discarded
Solution Approach 1:
The patent segments the attribution process into multiple discrete touchpoints along the user journey, assigning individual weights to each interaction rather than consolidating all credit to a single final touchpoint. This segmentation allows the system to evaluate and measure the contribution of each marketing impression separately, improving measurement precision while maintaining computational feasibility through systematic weight assignment.
Solution Approach 2:
The patent introduces a temporal dimension to attribution modeling by incorporating time decay functions that evaluate impressions based on their temporal proximity to conversion events. This adds a new dimension (time-weighted significance) to the traditional attribution approach, enabling more precise measurement of each touchpoint's contribution while maintaining model simplicity through the use of standardized decay curves.
2Measurement precision
If sophisticated multi-touch models are implemented to improve attribution accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent manages complexity by systematically varying key parameters such as time decay rates, touchpoint weightings, and conversion event types. Rather than creating complex structural models, the system achieves sophisticated attribution through parameter adjustment—modifying decay constants, weight distributions, and temporal windows—to adapt to different marketing scenarios while maintaining a relatively simple underlying computational framework.
Solution Approach 2:
The system handles complexity through configurable parameters that can be adjusted without restructuring the core model. By allowing flexible modification of decay rates, attribution windows, and touchpoint significance weights, the patent enables sophisticated attribution behavior through parameter tuning rather than through complex system architecture, thereby improving measurement precision without proportionally increasing device complexity.
3Measurement precision
If all preceding events are considered in attribution modeling, then measurement precision is improved, but loss of time increases due to processing larger datasets
Solution Approach 1:
The patent applies preliminary filtering and pre-computation techniques to reduce the dataset before full attribution analysis. By pre-identifying relevant touchpoints within defined temporal windows and pre-calculating decay weights for common time intervals, the system reduces the computational burden of processing all preceding events while maintaining complete attribution coverage. This preliminary action preserves measurement precision by retaining all relevant data points while significantly reducing processing time through selective optimization.
Data Source
AI summary
To implement a multi-touch attribution model, a conversion event resulting from user activity is programmatically detected. A set of events that precede the conversion event is identified. Respective events of the set occur on respective websites. The conversion event is attributed to multiple websites of the respective websites.


