Trust you can document, and a link profile strangers help write

Trust is a documentation problem before it is a ranking problem

Every website in every industry says that it can be trusted. The sentence costs nothing to type, which is precisely why it carries no information: a two-week-old shell company can write it as fluently as a forty-year-old family firm. What separates the two is never the claim itself but the paperwork standing behind it — whether there is a named human attached to the advice, a legal entity behind the invoice, a date on the page, and some way for a stranger to check any of it without having to ask you first.

So there are really only two kinds of site. One documents its trustworthiness and then steps back and lets the reader verify. The other asserts it in adjectives and quietly hopes nobody looks any closer. The VITON13 catalogue prices the first kind of work as E-E-A-T: pages of authors, expertise and trust, filed under the Trust category at $300 – $1 200, or 20 000 – 70 000 ₽, with a stated duration of ⏱ 1–3 days and a difficulty marked Experienced.

The unglamorous truth is that most of the raw material already exists inside the company. The founder really did spend a decade in the industry. The lead engineer really does hold the certification. The business really is registered somewhere, with a real address and a real phone that a real person answers. None of it appears on the site, because nobody was ever handed the job of writing it down. A trust build is mostly transcription, plus the editorial judgement to separate facts a sceptic can check from compliments that merely sound good.

What an author page has to contain before it counts as evidence

A headshot, a name and a job title is a byline, not an author page. It tells the reader who to blame; it does not tell them why to believe. Evidence has edges on it: the years someone worked and at which employers, the certification together with its issuing body and the date it was granted, the conference talk with the year attached, the repository or publication that lives on a domain you do not control. Each of those is a fact a suspicious reader can independently go and confirm.

The second requirement is attachment, and it is the one that gets skipped. An author page that no article links to is a cul-de-sac. The person has to be bound to the work: every piece they wrote carries their name, that name resolves to one canonical author URL rather than four near-duplicates, and the relationship is expressed in markup so that machines reading the page arrive at the same conclusion a human would. Consistency here is worth more than eloquence.

This is where the catalogue's ⏱ 1–3 days actually goes. Very little of it is writing. Most of it is sitting with the people who have the experience and extracting the specifics they consider too ordinary to mention, then chasing down the artefacts that prove those specifics, then wiring the whole set into the templates so it renders identically on every page. The difficulty rating of Experienced reflects the judgement involved, not the typing.

There is a blunt test for whether the page has crossed from decoration into evidence. Hand it to somebody who has never heard of your company and ask them to name one concrete thing this author has personally done. If they can only reply that the author is passionate about the field and has many years of experience, the page is still an assertion wearing a photograph. Rewrite it until the answer is a specific noun with a year next to it.

The company pages nobody reads until something goes wrong

Every site has a small cluster of pages with almost no traffic: the company page, the contact page, the terms, the refund policy, the privacy notice, the editorial policy if one exists at all. Analytics makes them look worthless. They are not read casually. They are read at exactly two moments — when somebody is about to spend money and wants to know who they are spending it with, and when something has gone wrong and they want to know what recourse they have.

What makes that cluster credible is aggressively boring specificity. A registered company name rather than a brand. A street address rather than a city. A phone number that connects to a person and an email that is not a contact form. Named people with named roles. Dates on the policies so a reader can see when they were last touched. None of that is persuasive writing; all of it is checkable, which is the entire point.

There is a second audience for these pages that most owners never think about. Anyone verifying a claim about you — a prospective partner, a journalist, a compliance officer, an automated system assembling a description of your business — lands here first, because this is where a site is supposed to state plainly what it is. A site that cannot answer the question who are you on its own pages leaves the answer to be improvised by whoever is asking.

How an anonymous site gets described by a system with nothing to quote

The catalogue entry for the trust work makes an unusual promise in its description: that it works both for rankings and for how the brand is described by AI models. That is not a fashionable add-on, it is a direct consequence of how such systems behave. A model reproduces what has been written down about you. When nothing specific has been written down, it does not invent specifics — it falls back on the category. You get described as a generic agency because generic is the only material available.

A named author with a verifiable history changes the material. It gives an attribution target: a person who said a thing, on a page with a date, backed by a credential that exists elsewhere. That is the shape of a statement a system can repeat with a source attached. Without it, your best insight circulates as an unattributed opinion, and unattributed opinions do not carry a brand name along with them.

None of this requires a separate product or a separate budget line. It is a by-product of documenting trust properly for human readers. The same author page that lets a buyer decide you are credible is the page that gives a machine something concrete to attribute. Build it once, for the sceptical human, and the second audience is served automatically.

Part of your link profile is written by people you have never met

Here is the asymmetry that makes link work different from everything else on a site. Your content is yours: you decide what goes on the page and you can change it this afternoon. Your link profile is not yours in the same sense. Anybody with a domain and ten minutes can point a link at you, and they do not need your permission, your knowledge, or any interest in your success. The profile is co-authored, and you only wrote part of it.

Most of the strangers are harmless. A forum comment, a scraper that mirrored your article, a directory nobody visits, an aggregator that republished a snippet. Some are worse: link farms that hoover up domains at random, and the occasional deliberate campaign aimed at making your profile look purchased. You did not choose any of it, which is exactly why the sane posture toward it is inspection rather than pride.

That reframes what an audit is. It is not a promotional exercise where you count your best links and feel good. It is defensive work — the same category as checking your locks or reading your own credit report. You are looking for what accumulated while you were not watching, and for what somebody else may have done in your name.

There is also archaeology involved. The catalogue description is direct about it: the profile is cleared of garbage and traces of past contractors. Anyone who has inherited an SEO account knows the pattern — three agencies over six years, each with a different theory of link building, none of whom left documentation, all of whose experiments are still attached to your domain and still counted.

What a link audit inspects, and why the band runs from $400 to $1 400

The catalogue lists this as Link profile audit and protection from other people's spam, in the Links category, at $400 – $1 400, or 20 000 – 60 000 ₽, with ⏱ 1–2 days and a difficulty of Experienced. The price line carries an unusual note for a catalogue: it states the market range for a link audit as $399–1 399, which tells you the quoted band is deliberately positioned against what the wider market charges rather than pulled out of the air.

What the work covers is a complete inventory before any opinion is offered. Every referring domain, not a sample. Each one classified: earned, bought, inherited, mirrored, automated, hostile. Then the harder pass, which is separating links that actually do something from links that merely exist, because a vast profile in which only a handful of domains carry any weight is a common and expensive misunderstanding.

The width of the band is not vagueness, it is a real difference in the size of the job. A young site with three hundred referring domains and one previous contractor is a different job from an established domain with tens of thousands of referring domains, two rebrands and a decade of undocumented experiments in the history. The ⏱ 1–2 days holds because the method is the same; what changes is the volume the method is applied to.

The deliverable that matters is two lists rather than a score. The first is what to keep and what to reproduce — the catalogue phrases it as seeing which links actually work and where to get similar ones, which turns a defensive audit into a source of targets. The second is what to disown, what to monitor, and what to leave alone because acting on it would cost more than ignoring it.

Early warning is the part of link work that cannot be done once

The service description ends with a clause that is easy to skim past: an early warning about someone else's link spam in your direction. That clause changes the shape of the engagement. Cleaning a profile is a state you reach on a particular Tuesday. The arrival of new links from strangers is a flow that does not stop because you finished an audit, and a flow cannot be handled by a one-time state change.

So a single audit is a photograph, and a photograph is only useful once you have a second one to compare it against. What you actually want to observe is the derivative — the rate at which unfamiliar referring domains appear. A slow trickle of odd links is background noise that every domain accumulates. Four hundred new referring domains in a week, from a neighbourhood you have never worked in, is an event, and the only way to see it as an event is to know what a normal week looks like.

In practice this means the first audit does double duty. It cleans up the accumulated mess, and it establishes the baseline that every later week is measured against. Without that baseline, a sudden influx is indistinguishable from good fortune, and by the time it is obviously not good fortune the cheap window for responding has usually closed. Baselines are dull to build and impossible to reconstruct after the fact.

Competitor analysis as a map of least resistance, not a scoreboard

The third service is Competitor analysis with monthly monitoring of changes, in the Competitors category, at $250 – $800, or 20 000 – 35 000 ₽, with ⏱ 1–2 days for initial analysis and a difficulty of Experienced. The description avoids the usual language of rivalry entirely. It promises a clear picture of how competitors collect traffic — which topics and pages work for them, which formats they use, and where you can go with the least resistance. That is a survey, not a contest.

That last phrase is the whole orientation. Most competitor reports are scoreboards: they rank everybody, hand you a number, and leave you with the useless conclusion that the market leader is ahead. A map of least resistance asks a different question. Not who is winning, but which specific ground is currently undefended, and which of those gaps you happen to be equipped to occupy without a two-year fight.

The initial analysis is where the map gets drawn. It identifies the topic clusters that carry a competitor's traffic and the formats they have settled on, which is usually more revealing than the topics themselves. It is also the natural moment to look at their trust surface — whether their articles carry named authors, what their company pages actually document — and at where their links come from, since the audit's list of reproducible sources and the competitor's list of sources tend to overlap.

Note the difficulty rating here, because it explains the price. Experienced is not about running the tools; the tools are cheap and anyone can export a list. It is about reading the export and correctly distinguishing a competitor's deliberate strategy from their accumulated accidents. Copying an accident is a common and entirely avoidable way to lose a quarter. Repetition across several months is what separates a plan from an accident, and a single export never shows repetition.

What actually changes between one month of tracking and the next

Monthly cadence is in the service name for a reason: a competitor's site is a moving object, and a report written in March describes a building that has since been renovated. Between any two months you will see new pages published, old pages quietly retired, navigation restructured, pricing edited, author pages appearing where there were none, and new referring domains that reveal where they have been spending effort.

The majority of that movement is noise, and the discipline is refusing to react to it. One new article is not a strategy. A price change may be a test. The signal is repetition with direction: three consecutive months of publishing into the same narrow cluster, a systematic rebuild of a section, a steady stream of links from one kind of source. Those are decisions somebody made and funded, and decisions are what you can plan against.

Here is an arithmetic example rather than a measurement of any engagement: the catalogue quotes ⏱ 1–2 days for initial analysis, and a monthly comparison is by nature a smaller job than building the first picture from nothing, since the baseline already exists. A reader can work out for themselves what twelve such comparisons would cost in working time at whatever rate they were quoted. The point of the exercise is that tracking is cheap relative to the first analysis, which is why the cadence is monthly rather than annual.

The other reason for a fixed cadence is behavioural. Competitor research done on impulse is done when somebody is anxious, which is the worst moment to interpret anything. A monthly report arriving on a schedule turns it into an ordinary input, read calmly, filed, and occasionally acted upon — rather than an emergency document read at midnight after a bad week.

Sequencing the three when the budget only covers one of them

The three services are not interchangeable, and understanding what each one does structurally makes the choice easier than comparing prices does. Documenting trust builds an asset that did not exist before. Auditing links protects an asset that already exists from damage that other people can cause. Competitor analysis does neither — it tells you where to point the next round of effort so that the asset you build is one somebody is looking for.

The catalogue bands are: E-E-A-T: pages of authors, expertise and trust at $300 – $1 200 or 20 000 – 70 000 ₽ over ⏱ 1–3 days; Link profile audit and protection from other people's spam at $400 – $1 400 or 20 000 – 60 000 ₽ over ⏱ 1–2 days; Competitor analysis with monthly monitoring of changes at $250 – $800 or 20 000 – 35 000 ₽ with ⏱ 1–2 days for initial analysis. All three carry the difficulty rating Experienced.

As an arithmetic example a reader can verify against those figures rather than a quoted package price: the three lower bounds add up to $950 and the three upper bounds to $3 400, which is the width of the decision you are actually making. In roubles the same sum runs from 60 000 ₽ to 165 000 ₽. Where you land inside it depends almost entirely on the size and history of the site, not on which options you tick.

If you must pick one, pick by symptom rather than by price. A site with no named humans on it should document trust first, because everything else amplifies a page that gives a stranger no reason to stay. A site that inherited an SEO account from a contractor whose methods nobody can describe should audit links first, since you cannot safely build on ground you have not surveyed. And a team that has budget to publish but no idea what to publish should buy the map before the construction.