Skip to content

Article 11 min read

Brand Authority: What Counts for Google When People Mention You Without a Link

by Daniel Sócrates published

Someone told you your company needs brand authority. Maybe it was your agency, a business partner, or a thirty-second video. That same week, three proposals arrived under that label.

One redoes your logo and your tone of voice. Another promises stories on news sites. The third talks about Google. They all cost money, and none of them explains what changes for the people searching for you.

Almost everything written on the subject is about perception. Customer trust, positioning, keeping what you promise in line with what you deliver. That matters, and it matters to people. Search works with a different material.

To the machine, brand authority is not what you say about yourself. It is what others write, in places you do not control.

This piece bridges the two conversations. And it ends with the decision you came here to make: what each of the three proposals actually buys.

The recommendation in the building group chat

Picture the group chat for your apartment building. A new neighbor asks if anyone knows a good electrician.

There is a flyer taped up in the elevator, with a photo, a slogan, and the line “the best in the area.” Nobody mentions the flyer. Three neighbors reply, at different times, with the same name. One says Marcus fixed his breaker panel. Another sends the number. The third just writes “Marcus, hands down.”

The new neighbor calls Marcus. The flyer said more, and said it with more polish. The call was decided by what three people with nothing to gain had written.

What others say about you, in places you do not control, on the subject you work in. In the guidelines Google writes for its quality raters, this is called reputation. 1

The instruction is direct. Raters are told to look for what outside, independent sources say about the site and about the people who create its content. And there is a tiebreaker. When a site says one thing about itself and reputable independent sources say another, the guidelines are blunt: “trust the independent sources.” 1

The same guidelines define authoritativeness. Raters weigh “the extent to which the content creator or the website is known as a go-to source for the topic.” 1 Known by people on the outside, like Marcus in the building chat.

One caution, so you leave without the most common myth in the industry. Raters do not touch anyone’s rankings. Google writes that “Search raters have no control over how pages rank.” It uses their work “as a restaurant might get feedback cards from diners,” to learn whether its systems seem to be working. 2 The guidelines describe what Google calls good. The measures the system calculates on its own have corroboration at their center, and they are explained there.

Where the machine reads a brand’s trail

Search only has access to what was written down and what was typed in. In that material, a brand’s trail shows up in three places.

On the page where Google explains how it ranks results, it gives an example of a quality factor. It looks at whether “other prominent websites link or refer to the content.” According to the page, this “is generally a good sign that the information is trustworthy.” 3

Read “or refer” again. The industry spent years counting only links. The official sentence puts mentions on the same line.

A patent granted to Google in 2014 separates the two with precision. On one side is the link you click. On the other is a reference or citation to a page, written on another page, with no link at all. 4 This is called an implied link.

The same patent carries the filter that matters to anyone buying media. The connection only counts when both ends are independent. The text lists what breaks independence: pages with the same owner, the same host, or the same creator. And pages with identical or very similar content. 4

Read that list with a mass-distributed press release in mind. The same text on forty news sites, on the same day, is identical content repeated forty times.

Now the limit, and I do not bend on it. A granted patent proves investment and engineering direction, and nothing more. A system running today sits outside what it can prove. 5 This one, in fact, shows up on Google Patents as expired for nonpayment of fees. 4 What it gives you is how Google engineers thought about the problem, with no weights attached.

The same patent counts a second number: searches that refer to the site. The document’s examples are a site’s name followed by “news” or “restaurant reviews.” And each person is counted only once. 4

In December 2024, evidence of a different kind surfaced. A researcher found a flaw in a Google system, and the company paid his team $13,337 for the discovery. 6 According to the technical coverage, the exposed data included a quality score per subdomain, from 0 to 1. 7

Two of the three inputs described are about the brand. The frequency of brand searches combined with other terms. And the presence of the site’s name in anchor text across the web. The third is the rate at which the site gets chosen in results, especially outside the top position. 7

Classify the evidence before you use it. This is Google’s own data, read by outsiders, in a 2024 snapshot. Nobody outside the company knows how much each input weighs. 8 Anyone promising you rankings in exchange for brand searches is selling a sentence the source never says.

Now bring this back to your business. People typing your company’s name together with a topic is the cheapest signal that your name gets around. 9

Your company’s record in Google’s catalog

The third place is the quietest. Since 2012, Google has kept a catalog of things in the world, each with an identifier, attributes, and relationships. 5 Your company may be in there as one specific thing. Or it may be just a name that looks like many others.

How does a company get into that registry? My reading, and I flag it as mine, has two parts. Independent sources state the same facts about you. And the system can conclude that they are all talking about the same subject. 10 How Google came to see companies and people as things is covered in entities and Google. The box with a logo and a description beside the results is the public display of that catalog. The knowledge panel has its own page here.

AI answers pull in the same direction. A September 2025 study compared AI search engines across several industries and languages. It found a strong preference for third-party sources over a brand’s own content and social media. 11 Marketing has an old name for those sources: earned media. The study is a preprint, not peer-reviewed, and I read it as an early signal.

Three proposals, three different purchases

Go back to the three proposals from the opening. With the trail in hand, you can tell what each one buys.

The branding agency delivers positioning, visual identity, and tone of voice. It is real work, and it speaks to people. To the machine, everything that comes out of it carries your signature. It is a declaration from an interested party: it helps identify who you are, without confirming anything on its own. 10 The watch point is the name. A change of name or description leaves everything already published about you on other people’s sites out of date. Before you approve it, ask who will update those sources. 10

The PR firm can produce exactly the trail the machine reads. It can also produce volume with no proof at all. The yardstick that separates the two fits in one question: who decided to publish? 9 In earned coverage, the outlet decided. In a mass-distributed release, a distribution algorithm decided. In a directory, or in material you upload yourself, you decided. 9

That is why the PR contract needs a new yardstick. Count stories bylined by a named reporter. Count distinct media groups. Count the pieces where your company is the subject of the reporting. And count how many carry your name in the headline. 9

The Entity SEO consulting firm works on the part the other two never touch. It settles the canonical version of the company’s name and facts. It publishes that version on an official home, at a fixed address. It aligns the sources that already exist and takes a snapshot of the before, so there can be an after. 9 The press is still needed, and now it has the right place to confirm.

The three add up. The real decision is who ties all three to the same name and the same facts.

Four questions for any proposal

Take these four to your next meeting, including one with us.

  1. What does today’s snapshot look like, with a date on it? A screenshot of a search for the name in an incognito window, plus what ChatGPT and Gemini say about the company. 5 9
  2. Who will decide to publish each mention you deliver?
  3. How many sources I do not control will start confirming the same facts about me?
  4. What changes in the name or description already published elsewhere?

Three offers end the conversation on the spot. A knowledge panel with a date in the contract, because nobody controls Google’s catalog from the outside. 9 10 A package placing you on dozens of news sites in a single day. That is the identical content the patent allows to be treated as non-independent. 4 And a Wikidata item created with nothing to back it up, because the deletion is public and ends up burning your second try.

The decision, with our stake disclosed

If your customers know your brand and almost nothing about it is written outside your site, start with the foundation. Canonical version, official home, aligned facts. 9 Then hire PR, measured by the four counts above. A disclosure, right at the point of recommendation: SEO Genome sells this foundation work, as part of our SEO consulting. The recommendation rests on the sources in this piece, which apply to any provider. Who you hire to do it is your call.

Before you answer any proposal, run the search Google’s guidelines teach its raters. Type your company’s name, a space, and -site: followed by your domain. Something like yourcompany -site:yourcompany.com. 1 What comes up is what is being said about you off your site. Profiles you run yourself may still show up, and those come off the list. 1

Open the top ten results and write down, next to each one, who decided to publish it. From today on, when someone offers you brand authority, you will ask where it will be written, and by whom. Take that sheet to the first meeting.

Frequently asked questions

Is brand authority the same as domain authority?

They are different things. Domain authority is a score created by an SEO tool. Moz, which owns the best-known one, puts it plainly: “DA is a predictive metric developed by Moz and is not used by Google to determine search engine rankings.” 12 Google’s internal documentation, leaked in 2024, has an attribute called siteAuthority. Nobody outside the company knows how much it weighs. 13 The tool’s number is an estimate made from the outside.

Does a strong social media presence count as brand authority?

It counts for the people who follow you. As evidence for the machine, it carries little weight, because it is you talking. The rater guidelines use this very example. The official social media pages a company maintains “would not be considered independent sources of reputation information about the company.” 1

Can a small business with no press coverage build this?

It can, and the guidelines themselves acknowledge the starting point. They say “small websites may have little or no reputation information,” and that this “is not indicative of high or low quality.” Many small local businesses rely on word of mouth. 1 The raw material is usually close by: your licensing board, the chamber of commerce, the insurance networks that list you, the supplier that mentions you. 10

How long does it take?

Nobody on the outside can say, and anyone who gives you a date is guessing. A relationship you declare takes an afternoon. A relationship third parties confirm takes years. 9 What can honestly be promised is the measurement: the same snapshot, under the same conditions, at 30, 60, and 90 days. 9

Notes and sources

  1. Google, “Search Quality Rater Guidelines” (General Guidelines), version of September 11, 2025, 182 pages, https://static.googleusercontent.com/media/guidelines.raterhub.com/en//searchqualityevaluatorguidelines.pdf. Accessed September 13, 2026. Section 2.5: “What do outside, independent sources say about them? When there is disagreement between what the website or content creators say about themselves and what reputable independent sources say, trust the independent sources.” Section 3.3: the heading “Reputation of the Website and Content Creators.” Section 3.3.3: the example search [ibm -site:ibm.com], described as “A search for IBM that excludes pages on ibm.com”, and the note that official social media pages the company maintains “would not be considered independent sources of reputation information about the company”. Section 3.3.5: “small websites may have little or no reputation information. This is not indicative of high or low quality. Many small local businesses or community organizations have a small ‘web presence’ and rely on word of mouth.” Section 3.4, Authoritativeness: “Consider the extent to which the content creator or the website is known as a go-to source for the topic.”
  2. Google Search Central, “Creating helpful, reliable, people-first content,” https://developers.google.com/search/docs/fundamentals/creating-helpful-content, page last updated December 10, 2025. Accessed September 13, 2026. Passage: “Search raters have no control over how pages rank. Rater data is not used directly in our ranking algorithms. Rather, we use them as a restaurant might get feedback cards from diners. The feedback helps us know if our systems seem to be working.”
  3. Google, “How Does Google Determine Ranking Results,” How Search Works, https://www.google.com/search/howsearchworks/how-search-works/ranking-results/. Accessed September 14, 2026. Passage: “one of the factors used to determine quality is understanding if other prominent websites link or refer to the content. This is generally a good sign that the information is trustworthy.”
  4. Patent US 8,682,892 B1, “Ranking search results,” inventors Navneet Panda and Vladimir Ofitserov, assignee Google LLC. Filed September 28, 2012, granted March 25, 2014. https://patents.google.com/patent/US8682892B1/en. Accessed September 13, 2026. Passages: “An implied link is a reference to a target resource, e.g., a citation to the target resource, which is included in a source resource but is not an express link to the target resource.”; resources “owned by the same entity, hosted by the same entity, or that were created by the same entity” and resources with “identical or similar content” can be considered non-independent; “a reference query for a group of resources is a search query that has been submitted to a search engine and has been classified as referring to a resource in the group”, with the examples “example sf news” and “esf restaurant reviews”, and a count of one query per unique user. Google Patents lists the patent as expired for nonpayment of fees (“Expired,” “Fee Related”), with the caveat that the legal status is the platform’s presumption.
  5. Daniel Sócrates, SEO de Entidade (published in Portuguese), 2026, chapter 3, “Sua empresa é uma entidade. O Google só precisa descobrir isso” (Your Company Is an Entity. Google Just Has to Find Out). Drawn from it: the catalog of things Google announced on May 16, 2012, with identifier, properties, and relationships; the rule of evidence for patents (a grant proves investment and engineering direction, without proving a system in production); and the day-zero snapshot, with the question asked of ChatGPT and Gemini.
  6. Danny Goodwin, “Exploit reveals how and why Google ranks content,” Search Engine Land, December 16, 2024, https://searchengineland.com/google-exploit-scoring-classifications-449333. Accessed September 13, 2026. Google “paid his team $13,337 for discovering the endpoint vulnerability”; among the factors in the quality score, “Brand visibility (e.g., branded searches, or searches that include the brand’s name)”.
  7. Luis Rijo, “Google’s site quality scoring system revealed through endpoint discovery,” PPC Land, December 19, 2024, https://ppc.land/googles-site-quality-scoring-system-revealed-through-endpoint-discovery/. Accessed September 13, 2026. The score “assigns values between 0 and 1 to websites at the subdomain level”, with three inputs: “Frequency of brand-specific searches combined with other search terms”, “Selection rate of the website in search results, particularly when not in the top position”, and “Prevalence of the website’s name or brand in anchor text across the web”.
  8. Daniel Sócrates, SEO de Entidade (published in Portuguese), 2026, chapter 7, “Confiança não é opinião. É cálculo” (Trust Is Not an Opinion. It Is a Calculation). Classifying the evidence by category, with the exploit as Google’s own data interpreted by third parties, a 2024 snapshot of one subsystem, with no known weight for any input.
  9. Daniel Sócrates, SEO de Entidade (published in Portuguese), 2026, chapter 12, “Dê à sua entidade uma casa oficial” (Give Your Entity an Official Home). Drawn from it: brand search as the cheapest signal that a name gets around; the hierarchy of mentions by the question “who decided to publish”; the four changes of target in the PR contract; the canonical version and the entity home; the protocol of screenshotting the search for the name at 30, 60, and 90 days; the difference between a declared relationship, which takes an afternoon, and a confirmed relationship, which takes years; and the warning about anyone selling a knowledge panel with a date in the contract.
  10. Daniel Sócrates, SEO de Entidade (published in Portuguese), 2026, chapter 4, “O cartório do Google: como strings viram things” (The Google Registry: How Strings Become Things). Drawn from it: entry into the registry through independent sources stating the same facts, a criterion the book flags as the author’s inference, supported by Google Cloud’s reconciliation documentation and the Knowledge Vault paper (KDD 2014); the declaration from an interested party, which helps identify without confirming on its own; the question about older sources when a new name is proposed; the sources within reach of a neighborhood clinic; and the absence of any channel or timeline for getting into the Knowledge Graph.
  11. Mahe Chen, Xiaoxuan Wang, Kaiwen Chen, and Nick Koudas, “Generative Engine Optimization: How to Dominate AI Search,” arXiv 2509.08919, submitted September 10, 2025, preprint not peer-reviewed, https://arxiv.org/abs/2509.08919. Accessed September 13, 2026. Passage: “AI Search exhibit a systematic and overwhelming bias towards Earned media (third-party, authoritative sources) over Brand-owned and Social content, a stark contrast to Google’s more balanced mix.”
  12. Moz, “Domain Authority: What is it and how is it calculated,” updated March 18, 2025, https://moz.com/learn/seo/domain-authority. Accessed September 13, 2026. Passage: “DA is a predictive metric developed by Moz and is not used by Google to determine search engine rankings.”
  13. Google’s Content Warehouse API documentation, leaked in 2024, module CompressedQualitySignals, in the public mirror of the google_api_content_warehouse v0.4.0 package, https://hexdocs.pm/google_api_content_warehouse/0.4.0/GoogleApi.ContentWarehouse.V1.Model.CompressedQualitySignals.html. Accessed September 13, 2026. Attribute siteAuthority, described as “converted from quality_nsr.SiteAuthority, applied in Qstar”. Google confirmed the material was authentic in Mia Sato, “Google confirms the leaked Search documents are real,” The Verge, May 29, 2024, https://www.theverge.com/2024/5/29/24167407/google-search-algorithm-documents-leak-confirmation. Accessed September 13, 2026. The material documents what is stored. The weight of each attribute is outside it.

Want to see SEO Genome more often on Google? Add it as a preferred source.

Daniel Sócrates

Founder and CEO of SEO Genome

Author of AI Search Optimization. Writes about what the company applies in client work.

danielsocrates.com.br

Let us look at what the machine understands about your company

A diagnostic conversation, with no slide deck. You bring your website address. We bring what Google and the AI assistants already know about it today.

Talk to us on WhatsApp