Machine Relations definition: Earned Authority is an MR label for independently published references to an entity.
Machine Relations definition: Earned Authority is an MR label for independently published references to an entity. The reference may be an editorial article, attributed expert commentary, an analyst report, an academic paper, or an institutional publication.
Premise: Machine Relations records these references separately from material published by the entity itself. The separation describes who controls the publishing surface. It does not establish that a reference caused an AI citation, that an answer engine trusted it, or that one publication carries a universal weight.
Premise: the distinction matters because AI answer engines — including ChatGPT, Perplexity, Google AI Overviews, and Gemini — retrieve and cite from a mix of sources when generating answers. An independently published reference occupies a different position in that information environment than a page the entity itself controls. G2's 2026 AI Search Insight Report found that 85% of B2B software buyers perceive vendors cited by AI more favorably, and 69% chose a different vendor than initially expected based on AI recommendations (G2, 2026). These observations describe buyer behavior around AI-surfaced citations, not a universal rule that independent references cause AI citation.
For each observed reference, record:
Premise: Keep the original item and its syndicated copies in one source family while retaining the URLs of the copies. This prevents a distribution count from being presented as a count of independent editorial decisions.
Premise: The Earned Authority field includes independently controlled editorial, analyst, academic, government, and institutional references. Brand-owned articles, social posts, paid placements, and items whose editorial status is unclear stay in separate fields. These are MR classification rules, not claims about how an answer engine ranks or selects sources.
Premise: the classification reflects a structural property of the information environment. Research by Ipsos, Muck Rack, and VaynerX found that 57% of consumers identify traditional news as their most trusted source, while 85% of adults aged 18–34 get their news from social platforms (Ipsos, 2025). The Machine Relations protocol does not assign trust values based on these consumer preferences; it records the publishing control of each source so that analyses can separate editorial decisions from distribution mechanics.
Premise: the following table describes the protocol's classification rules. It does not assign a weight, trust level, or AI-citation probability to any source type.
| Source type | Publishing control | MR classification |
|---|---|---|
| Independent editorial | Publisher controls | Earned Authority |
| Brand-owned content | Entity controls | Owned |
| Sponsored placement | Advertiser pays, publisher hosts | Paid / disclosed |
| Syndicated copy | Original publisher controls; distributor reproduces | Source family of origin |
| Unclear editorial status | Cannot be determined | Unclassified |
Premise: Earned Authority is not brand awareness, backlink volume, total mention volume, search ranking, AI citation rate, or business impact. Those observations require their own denominators and evidence.
Premise: A higher reference count does not by itself prove greater trust, better visibility, stronger citation performance, or a causal effect. Comparisons should disclose the query set, engine set, source-classification rules, and observation window used for the analysis.
Premise: the concept is deliberately narrow. It answers one question — how many independently published references to this entity exist in the declared observation window? — and separates that count from every downstream interpretation.
Mixing owned and independent material. Premise: the count becomes ambiguous when the protocol does not identify who controls each publishing surface.
Counting syndication as independent origin. Premise: copies can be recorded without treating each copy as a separate editorial decision.
Changing the denominator. Premise: results from different source sets or observation windows should not be presented as one continuous measure.
Inferring causality from co-occurrence. Premise: a reference and an AI citation can appear in the same observation period without proving that one produced the other.
Assigning universal publication tiers. Premise: the protocol records the actual publisher and observed citation behavior instead of assuming a fixed trust value for a publication name.
Premise: The MR Stack places Earned Authority at Layer 1 as an analytical ordering choice. Entity Clarity, Citation Architecture, distribution, and measurement remain separate layers with separate observations. The layer order is a Machine Relations model, not a measured sequence of causation.
Premise: within the stack, Earned Authority establishes a baseline: what independently published material exists about an entity before any AI citation or search observation is evaluated? The other layers — Entity Clarity, Citation Architecture, distribution, and measurement — each add a separate observable dimension. No single layer, including Earned Authority, determines an outcome by itself.
Machine Relations definition: Earned Authority is an MR label for independently published references to an entity.
Premise: Record every observed reference, then classify its publishing control and editorial status. Only independently published references enter the Earned Authority field; ambiguous items remain visible in a separate field rather than being forced into the count.
No. Premise: It records independently published references. Trust, citation selection, and business effects require separate evidence.
No. Premise: Earned Authority describes a source set. AI citation rate describes observed citations across a declared query and engine panel. Report the two separately before testing whether they move together.
Premise: the separation isolates who controlled each publishing decision. An entity can publish its own material, pay for placement, or be referenced independently. Each path represents a different editorial decision process. The MR protocol keeps them distinct so that analyses do not conflate distribution volume with independent editorial coverage.
Premise: Earned Authority records one input to the information environment that AI engines retrieve from. Whether an answer engine cites a particular independent reference depends on the engine's retrieval and citation selection process — which is not exposed publicly and varies by product. An Ahrefs study of 863,000 keywords found that only 38% of pages cited in Google AI Overviews also appeared in the top 10 organic results — down from 76% in mid-2025 (Search Engine Journal, 2026). This divergence illustrates why Earned Authority and AI citation presence are separate observations: a page's search rank does not predict its citation status. The protocol records the reference; measurement records the citation; analysis tests whether they correspond.
Machine Relations definition. This is an established industry term. This page is Machine Relations' definition of it — not a claim to have originated the term.
Machine Relations' own methodology, dataset, and research pages related to this term. These are self-references, listed separately from Sources — they are not independent evidence.
Related concepts
Supporting research
Framework context