Privacy-Preserving Consumer Scoring via Data Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

There is a need to generate behavioral scores for consumers based on transaction data while maintaining their privacy, as existing methods often compromise personal information and violate regulations.

Innovation Solution

A system and method involving multiple computing systems where consumer characteristics are encrypted and disguised to prevent personal identification, allowing for the generation of behavior prediction scores without revealing personally identifiable information, by using one-way encryption and mapping encrypted account identifiers across systems.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If entities gather detailed consumer information including transaction data, then they can better target advertisements and understand consumer behavior, but consumer privacy is compromised and regulatory requirements are violated

Engineering Contradiction:
Improveconsumer behavior informationVSAvoidprivacy violation
Core Design Contradiction:
Loss of informationVSObject-affected harmful factors

Solution Approach 1:

The system segments consumer data processing across multiple computing systems: a first computing system holds encrypted account identifiers and consumer characteristics, while a second computing system holds transaction data. This segmentation ensures no single system possesses both personally identifiable information and transaction data, preventing privacy violations while enabling behavioral analysis through coordinated processing.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces encrypted account identifiers as an intermediary element that bridges the gap between consumer identification and transaction data. These encrypted identifiers allow the system to link consumer characteristics with transaction behavior without exposing actual consumer identities, thus maintaining privacy while enabling targeted advertising and behavioral scoring.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If consumer characteristics are kept in plain form for accurate scoring, then behavior prediction accuracy is improved, but personally identifiable information is exposed

Engineering Contradiction:
Improvebehavior prediction accuracyVSAvoidpersonal information exposure
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The system changes the parameter state of consumer characteristics by applying encryption transformations. Consumer characteristics are stored and processed in encrypted forms (e.g., hashed account identifiers, encrypted demographic data) that preserve the statistical and behavioral patterns needed for accurate scoring while eliminating personally identifiable information. The encryption parameters are chosen to maintain data utility for behavioral analysis while ensuring privacy protection.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3170138B1Method and system for maintaining privacy in scoring of consumer spending behavior
Publication Date: 2021.03.03 MASTERCARD INT INC
  • EP3170138B1 patent drawingFigure 1
  • EP3170138B1 patent drawingFigure 2
  • EP3170138B1 patent drawingFigure 3A

AI summary

A method for maintaining consumer privacy in behavioral scoring includes a first computing system and a second computing system. The first computing system disguises consumer characteristics and maps disguised consumer characteristics to unencrypted account identifiers, and then transmits the data to the second computing system. The second computing system encrypts the account identifiers upon receipt, and maps the encrypted account identifiers to anonymous transaction data. The second computing system uses the transaction data to calculate consumer behavioral scores, and then generates a scoring algorithm that uses disguised consumer characteristics to calculate consumer behavior scores based on the calculated consumer behavioral scores and corresponding disguised consumer characteristics. The generated algorithm is then returned to the first computing system, with the second computing system not receiving any unencrypted account identifiers, any undisguised consumer characteristics, or any personally identifiable information