Multi-touch attribution model with energy decay weighting

Resolve Bottlenecks,
Find Innovative Solutions
Generate 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

VSEngineering 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

Engineering Contradiction:
Improveease of computation and explanationVSAvoidaccuracy of attribution
Core Design Contradiction:
Ease of operationVSMeasurement precision

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.

Inventive Principle:
Principle #1Segmentation

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.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If sophisticated multi-touch models are implemented to improve attribution accuracy, then measurement precision is improved, but device complexity increases

Engineering Contradiction:
Improveaccuracy of attributionVSAvoidcomplexity of modeling system
Core Design Contradiction:
Measurement precisionVSDevice complexity

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.

Inventive Principle:
Principle #35Parameter changes

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.

Inventive Principle:
Principle #35Parameter changes

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

Engineering Contradiction:
Improvecompleteness of attribution dataVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

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.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS10649818B2Multi-touch attribution model for valuing impressions and other online activities
Publication Date: 2020.05.12 AGGREGATE KNOWLEDGE LLC
  • US10649818B2 patent drawing
  • US10649818B2 patent drawing
  • US10649818B2 patent drawing

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.