Link analysis is not your address book

Link analysis, in How Oernoe Search works, is a ranking factor about the public web. The guide defines it in plain language: how many other reputable websites link to the page, and the idea that more citations from authoritative sources indicate higher quality content. That is a statement about documents citing documents. It is not a statement about your contacts, your phone, your chat roster, or any personal graph hanging off an account.

This essay exists because “links” is an overloaded word. In ordinary speech, a link can mean a URL on a page, a social connection, a saved contact, or a follow edge in an app. In ranking copy, link analysis means the citation graph of public sites. Collapsing those meanings turns a page-quality signal into a story about people. We refuse that collapse.

## What the ranking list actually contains

Privacy-preserving ranking factors on the guide include content relevance, authority and expertise, link analysis, page freshness, aggregated user satisfaction signals, mobile-friendliness, and page speed. The guide then says what is not on the list: your search history, your location, your device type, your age, your interests, or other personal data used as an advertising profile of the querier.

Link analysis sits with the document signals. It asks whether other reputable sites point at the candidate page. Authority and expertise ask whether the hosting site looks trustworthy. Content relevance asks whether the page addresses the query. Freshness asks how recently the page was updated, which matters more for newsy topics. Mobile-friendliness and page speed ask how the document behaves when served. Satisfaction signals, when present, are aggregated and anonymized at scale.

None of those bullets open your address book. None of them require syncing a contact list into Search. None of them score you as a person against your friends. The algorithm is trying to decide which public pages answer a phrase. Citations between sites are one clue about which pages other publishers thought were worth pointing to.

## Citations are public, contacts are not the ranking object

A citation on the open web is visible without logging into your account. Another site published a hyperlink. Crawlers that respect robots rules can see that edge when they fetch the citing page. The edge becomes part of how pages relate to each other in the index.

Your address book is a different object. It lives behind authentication. It is about people you chose to store, not about publishers who chose to cite a URL. Mixing those graphs would be a product decision we do not make, and a privacy failure we do not want. Search queries are not sold to build advertising profiles. Ranking is not allowed to quietly become a social graph product under a technical synonym.

When people say “who links to me,” they sometimes mean “which sites cite my article,” which is fair ranking talk. Sometimes they mean “which humans in my circle mentioned me,” which is not what the guide’s link analysis bullet describes. If you need the human circle, that is a different product family entirely, and it is not how we explain Search ranking.

The difference shows up in what a crawler is allowed to see. Public HTML with an href is fair game under robots rules the publisher chose. A private contact record is not sitting in that HTML. Treating both as “links” for ranking is a category error dressed up as cleverness. We would rather sound pedantic than sound like we ingest your contacts to decide which gasket page ranks higher.

## Why reputable citations matter without becoming a people score

Not every inbound link is equal in the sentence we published. The guide says reputable websites and authoritative sources. That is quality language about publishers, not a popularity contest counted in raw edge volume alone. A cluster of low-quality mirrors pointing at each other is not the same as independent editorial sites citing a primary document.

Still, even a careful citation graph is about pages. A page can earn citations because it is useful, clear, or primary. A person can be mentioned on many pages without those mentions being the same as account-level friend edges. Ranking looks at the page in the index. It does not need your contact list to decide whether that page matches “replacement gasket torque spec” or “county clerk hours.”

Authority signals are related and still distinct. Authority and expertise ask how trustworthy the site is. Link analysis asks about citations to the page. You can write essays about either without turning either into a profile of the searcher. This piece stays on link analysis as citation quality on the public web graph.

Reputable citations also fail gracefully when they are absent. A brand-new primary document can be the right answer before anyone has cited it. Relevance can still surface it. Continuous crawl can still discover it through other paths. Link analysis is one factor among hundreds of signals the guide summarizes, not a veto that says “no citations, no rank.” Inflating it into a social score would make that failure mode worse, because new pages and new people would be punished for lacking a contact graph they should never have been asked to provide.

## Distinct neighboring essays, deliberately not this one

Authority signals are not your history. Ranking signals are about pages, not people. Content relevance is query to page. Page freshness is not a profile. Those lines are neighboring clarifications. They share a theme: do not personalize the ranking story into an advertising file of the human who typed the query.

This essay’s job is narrower. It only separates link analysis from address-book thinking. If you need the relevance definition, open the relevance essay. If you need freshness, open freshness. Do not paste all of them into one mega-claim and call it a ranking explainer. Readers who skim five titles into one paragraph are the reason we keep publishing these separations.

## How equal results fit beside citation signals

The guide says the algorithm works the same way for every person unless custom preferences were saved on the account. Two friends searching the same pizza phrase get the same results under that default. Link analysis still runs. It runs on pages, not on which friend has more contacts. Personalization that exists is explicit: saved searches, bookmarked sources, language and region preferences, filters you configure. Those are under your control. They are not an automatic harvest of your social graph for ranking.

That design keeps citation quality useful without turning Search into a who-you-know machine. The citation graph can still prefer a well-cited municipal page over a scraped mirror. The querier’s contact list never had to enter the scoring function for that preference to exist.

Equal treatment also protects the meaning of a citation. If results shifted because one querier followed more publishers on a social app, “reputable citation” would quietly mean “popular in your circle.” That is not what the guide says. The guide points at websites linking to a page. Keeping that definition stable is part of keeping Search from becoming a personalized advertising profile machine under a softer name.

## What link analysis is not for

It is not a way to import your phone contacts into Search. It is not a trust score on you as a human. It is not a back door into Health notes, Chat threads, or Drive files. Search listings for people and companies are a separate product path with an approval step; they are not the same thing as inbound hyperlinks from other sites to a URL. Confusing listing objects with citation edges is another mash we will refuse in print.

It is also not a promise that more links always win. The guide’s sentence is about reputable citations as a quality indicator. Spammy link schemes are an old web problem. We are not going to publish a cookbook for gaming citations. We are going to keep saying the signal is about public page quality.

Support tickets that attach a spreadsheet of names and ask us to “count these as links” will get a short answer: those are not the ranking object. Publish on the open web if you want public citations. Use account preferences if you want your own filters. Do not ask ranking to become a CRM.

## Publisher honesty around ads and Search

Anoepal operates both Oernoe Search and the publisher site on www.oernoe.com. Sharing an operator does not authorize a shared advertising file of your queries. Selected finished publisher pages may load Google AdSense. Search does not sell queries for advertisers’ marketing. Saying link analysis is about public citations does not require us to pretend the publisher hostname never shows ads. We disclose ads where they appear. We will not claim the whole network has no advertising systems.

Journal writing here has to stay about behaviors we run. Link analysis as defined on the ranking list is one of them. Address-book ranking is not. If an older promotional URL implied otherwise, treat the live How Search works page as the authority and leave the stub behind. What we publish already pushed thin slogan pages out of the public index for a reason. This clarification belongs in the index because it names a real factor and draws a real boundary.

## Practical reading for site owners

If you want citation quality to reflect real editorial attention, publish something worth citing and earn links the ordinary way: other sites choosing to point at you. Keep your pages fetchable under robots rules you intend. Make the content match the queries you hope to answer. That is relevance plus honest discovery. It is not a request that we read your CRM.

If your page is missing, the problem may be crawl and discovery, not link analysis. If your page is present but ranks below a clearer competitor, look at relevance and document quality first. Dumping a contact export into a support ticket will not change either diagnosis.

Site owners also ask whether a people or company listing boosts link analysis. It does not work that way in the sentence we published. A listing is a submitted, approved Search object. A citation is a hyperlink from another reputable site to a page. You can have one without the other. You should not expect the approval queue for listings to manufacture citation equity on the open web. Those are separate levers, and only one of them is “link analysis” on the ranking list.

## Closing

Link analysis on our ranking list means citations from other reputable sites to the page. It is about public web graph quality. It is not your address book, not your personal graph, and not a license to score people instead of pages. Keep the overloaded word “link” tied to the guide’s definition when you read us. If someone sells you ranking as a social product in disguise, ask them to quote the bullet. Ours points at websites citing a page, not at humans stored in a contacts file.