Methodology

How We Measure
Authority

We map publicly available data and the connectivity between it, across tens of thousands of topics, to determine who genuinely holds authority in any field and how fast that authority is growing.

Data Advantage

A Data Asset Years in the Making

We have maintained continuous, compliant API access to the major social platforms for years: a persistent programmatic relationship with real historical depth that most newer entrants simply cannot recreate.

The cross-platform identity graph we have built links social profiles, video channels, podcast appearances, and editorial mentions into unified entity profiles. Every additional year of signal compounds: a profile enriched over years of activity carries signal quality that one assembled last month cannot match.

Every new data source we integrate creates new cross-references across every existing profile, compounding the value of what is already collected. Recreating this today means navigating API restrictions, platform policy changes, and historical data gaps that did not exist when we started, regardless of compute budget.

  • Social media presence across major platforms
  • Podcasts and long-form audio presence
  • Editorial news coverage and press mentions
Multi-Source Ingestion
Social media presence, podcasts, and news coverage pulled via compliant API access.
Identity Resolution
Cross-platform profiles linked into a single entity, regardless of how many surfaces a person appears on.
Signal Extraction
Interactions, citations, and mentions extracted and weighted by source quality and topic relevancy.
Authority Score
A normalized score and signal growth rate output for every indexed entity, updated continuously.
ENTITYPROFILEYOUTUBELINKEDINNEWSPRESSPODCASTSACADEMIC
One unified profile from five independent sources
Identity Graph

One Person, Many Sources. One Score.

An authority rarely exists on a single platform. A climate researcher may publish academic papers, host a podcast, maintain a YouTube channel, appear in news coverage, and hold an active LinkedIn presence. Each surface contains a fragment of the full picture.

Our identity resolution layer links those fragments into a single unified entity profile. The result is a score that reflects the whole person, not just whatever corner of the internet a query happens to surface first.

Critically, every new data source we add does not just enrich new profiles: it creates new cross-references across every entity already in the index. Five platforms is not five times the data, it is the connections between those five sources for every entity, compounding with each addition. That is why the index becomes harder to replicate with every platform we integrate, not easier.

Topic Coverage

Tens of Thousands of Topics, Structured

We do not score authority in a vacuum. Every entity is evaluated within a structured topic taxonomy that spans tens of thousands of fields, niches, and sub-disciplines. Topic relevancy is derived from the content an entity produces and the interactions that the entity receives.

Topic density matters: an entity that is highly connected within a dense, well-indexed topic cluster carries a different signal than one operating in a sparse niche. Both are scored, but the context shapes how signals are interpreted.

  • Structured taxonomy across tens of thousands of topics
  • Topic relevancy derived from content and connected entities
  • Topic density used to calibrate signal weight
  • Each entity scored within its most relevant topic context
Topic Classification
Topic relevancy confirmed
Matched
Connected entities in same topic
Confirmed
Topic density: high
Indexed
Topic relevancy unclear
Unscored
Entity excluded from topic
Taxonomy Lifecycle
AI safety
Emerging
Prompt engineering
Densifying
mRNA therapeutics
Indexed
Sparse legacy niche
Merged
Topic Taxonomy

A Taxonomy Built Bottom-Up

Our taxonomy is bottom-up. Topics are derived from the content and connection patterns observed in the graph itself, then curated by topic experts.

New topics emerge as clusters become dense enough to warrant separation; existing topics merge when their connection patterns collapse. This means emerging fields like AI safety, prompt engineering, and mRNA therapeutics appear in the index as soon as they have signal, rather than waiting for a static taxonomy to be updated.

9.47.88.1
High authority entityMid-tier entity
Scoring Model

Connectivity Determines Score

Authority is not self-declared. It is derived from how an entity is connected to others within its topic. The more interactions an entity has with relevant, high-authority sources, the stronger its score. Low-quality connections carry little weight; well-connected, relevant sources compound it.

This means follower counts and surface-level engagement do not drive the score. What matters is the quality and relevancy of the entities in your connection graph, and whether those connections are meaningful within your topic context.

Source quality is derived from the authority score of connected entities, calibrated by topic relevancy and the density of the topic cluster they belong to.

Cold Start

Bootstrapping the Graph

Every connectivity-based ranking system faces the same bootstrap problem: authority is derived from links to other authorities, so where does the first authority come from?

Our answer is a multi-source seeding process per topic, combining academic citation graphs, established editorial output, and cross-platform identity persistence.

Seed sets are not static: they are continuously audited and re-validated as the graph evolves, with decay applied to entities that lose relevant connections and growth signals applied to new entrants.

Academic Citation Graphs
Peer-reviewed citation networks provide an initial authority prior per topic.
Established Editorial Output
Vetted publications and bylines seed authorities with a known track record.
Cross-Platform Identity Persistence
Consistent presence across independent sources anchors a seed to a real entity.
Continuous Audit & Re-Validation
Seeds decay or grow as the graph evolves. No entry stays seeded on reputation alone.
Authority Score Over Time
t₀T0.50.81.0
Score growth as relevant connections accumulate
Two Outputs

Authority Score and Signal Growth

Every indexed entity receives two outputs. The authority score is a normalized measure of how well-connected and relevant an entity is within its topic, based on the quality and density of its connections to other indexed entities.

Signal growth tracks the rate at which that score is changing. A rising signal growth indicates that an entity is gaining new, relevant interactions faster than the topic average, making them a leading indicator of where authority in a field is shifting.

  • Authority score: a normalized connectivity and relevancy measure
  • Signal growth: rate of change relative to topic average
  • Rising signal growth flags emerging authorities early
Integrity

Is the Score Gameable?

Authority scoring is gameable in principle, like any ranking system. Three structural defenses make it expensive in practice:

  • Connectivity weighting: a new account's score depends on whether high-authority entities engage with it, not on its own activity. Buying followers does not move the score.
  • Cross-platform requirement: a single-platform stunt does not produce a unified profile, since identity resolution requires consistent presence across at least two independent sources.
  • Signal-growth tracking: anomalous spikes are flagged for review rather than rewarded.

We do not claim the system is unfakeable. We claim that faking it costs more than it returns.

Gaming Attempt
Bought 50k followers
No effect
Single-platform spike
No unified profile
Coordinated mention burst
Flagged for review
Score unchanged. Cost exceeds return.

Access the Authority Index

Query authority scores for any indexed entity via our API.

Get API Access