47 Signals: How Public Data Becomes Voter Intelligence
Every persuadability score in Political Leverage is built from publicly available data. No purchased consumer data. No social media scraping. No privacy violations. Here's exactly how it works.
The Public Record Foundation
Political Leverage's intelligence engine operates exclusively on data that is a matter of public record. This is not a philosophical choice. It is a legal and operational one. Public record data is defensible under campaign finance law, does not require consumer consent frameworks, and is available consistently across all 50 states.
The six primary data categories are: voter registration files (available from every state's Secretary of State office), FEC donation records (publicly searchable federal campaign contribution data), state and local campaign finance records, petition signature databases (public records in most states), property records (county assessor data available in all 50 states), and precinct-level election results (historical voting data published by state election boards).
Each data category provides a different lens on voter behavior. Voter registration files show party affiliation and voting frequency. FEC records show where voters put their money. Petition signatures show which issues voters care enough about to take public action. Property records correlate with economic sensitivity. Precinct data shows community-level voting trends.
Signal Extraction and Weighting
Raw public data becomes actionable intelligence through a three-stage process: extraction, normalization, and composite scoring. In the extraction phase, the system identifies specific behavioral indicators within each data source. A cross-party donation is not just a donation record. It is a signal that the voter's financial behavior contradicts their party registration.
The normalization phase adjusts signal strength based on recency, frequency, and context. A cross-party donation made in 2024 carries more weight than one made in 2016. A voter who has signed three petitions on healthcare is a stronger signal than one who signed a single petition. A precinct that swung 8 points carries more weight than one that swung 2 points.
The composite scoring phase combines normalized signals into a single persuadability score on a 0-100 scale. The scoring algorithm is not a black box. Every score includes a signal breakdown showing which data points contributed to the score and how much weight each signal carried. Campaign managers can inspect any score and understand exactly why a voter was flagged as persuadable.
Issue-Level Sensitivity Mapping
Persuadability is not a single dimension. A voter who is persuadable on healthcare may be completely decided on gun policy. Political Leverage's scoring system generates issue-level sensitivity scores in addition to the overall persuadability score.
Issue sensitivity is derived primarily from petition signature data, donation patterns to issue-specific PACs, and precinct-level ballot measure results. When a voter has signed a healthcare petition, donated to a healthcare-focused PAC, and lives in a precinct where a healthcare ballot measure passed by a narrow margin, the system assigns a high healthcare sensitivity score.
This granularity allows campaigns to match messages to voters at the issue level. Instead of sending a generic persuasion message, a campaign can send a healthcare message to voters with high healthcare sensitivity and an economic message to voters with high economic sensitivity. The result is higher relevance and higher conversion.
Validation and Accuracy
The value of any predictive model depends on its accuracy. Political Leverage validates persuadability scores against actual election outcomes using a holdout methodology. In each cycle, a random subset of scored voters is excluded from outreach and tracked through the election. The correlation between predicted persuadability and actual vote behavior provides a continuous accuracy benchmark.
In the 2024 cycle, the system achieved an 82% accuracy rate in predicting which voters would deviate from their party's candidate. This compares favorably to traditional partisan modeling, which typically achieves 65-70% accuracy on the same metric. The improvement comes from the behavioral signal layer that traditional models lack.
Accuracy improves over time as the system processes more election cycles and refines signal weights. The 2026 cycle will benefit from two additional years of donation data, petition activity, and precinct-level results that were not available when the system was first deployed.
Privacy and Compliance
Every data source used by Political Leverage is a matter of public record, available to any citizen or organization that requests it. The system does not purchase consumer data from data brokers, does not scrape social media profiles, and does not use any data that requires individual consent under state or federal privacy laws.
This approach is not just legally defensible. It is operationally superior. Public record data is consistent, verifiable, and does not depend on third-party data agreements that can be revoked. When a campaign uses Political Leverage, they can explain exactly where every data point came from and why it is legally permissible to use it.
The system is reviewed quarterly by outside counsel specializing in campaign finance and data privacy law. Every new data source is vetted for legal compliance before integration. This is not an afterthought. It is a core design principle.