Stateful Data Corruption Prevention via Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems using stateful data, such as Internet cookies, to determine conversions based on impressions are susceptible to corruption, leading to inaccurate incrementality measurements as the number of impressions increases over time.
Innovation Solution
The use of specific data structures with fields for populating information and data normalization sets with functions applied to test and control data sets to remove corruption, ensuring incrementality is determined from uncorrupted impressions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional stateful data structures (e.g., Internet cookies) are used to track impressions, then the system can store and retrieve interaction information, but the data becomes corrupted over time leading to inaccurate incrementality measurements
Solution Approach 1:
The patent segments stateful data into multiple separate data structures, each tracking specific impression metrics independently. This segmentation prevents corruption from affecting the entire data set and allows for more reliable reconstruction of impression information even when individual segments become corrupted.
Solution Approach 2:
The patent implements preliminary validation and verification mechanisms that check data integrity before corruption can occur. By performing these checks in advance and using redundant storage of critical impression information, the system prevents corruption from compromising measurement accuracy.
2Productivity
If the number of impressions increases over time, then more conversion data is collected, but the corruption of stateful data increases leading to less accurate measurements
Solution Approach 1:
The patent implements feedback mechanisms that continuously monitor data quality and detection accuracy. When corruption is detected or anticipated based on impression volume thresholds, the system adjusts its tracking methods or triggers data validation routines to maintain measurement precision despite increasing data volume.
Solution Approach 2:
The patent changes parameters of the data structures used for tracking, such as implementing versioned cookie formats or adjusting data retention policies based on impression volume. These parameter changes allow the system to maintain accuracy as the scale of operations increases.
3Adaptability or versatility
If stateful data is used to determine conversions, then conversion tracking is enabled, but incrementality determination becomes unreliable due to data corruption
Solution Approach 1:
The patent introduces intermediary data structures and validation layers that mediate between the raw stateful data and the incrementality calculation process. These intermediaries filter out corrupted data and ensure that only reliable information contributes to incrementality determinations, maintaining both tracking capability and reliability.
Data Source
AI summary
Methods and systems described herein recite the use of specific data structures for stateful data that include fields for populating with specific information that may be used to prevent and/or minimize the corruption in stateful data. The systems and methods may recite the application of specific data normalization sets that include functions to be applied to test data sets and control data sets, respectively, in order to prevent and/or minimize the corruption. Finally, the methods and systems may recite the determination of incrementality being based on uncorrupted impressions corresponding to test data sets and control data sets, respectively.