Trust doesn't exist online. Not the way we talk about it offline.
When you walk into a shop and the owner remembers your name, trust is being built the slow way. Presence, repetition, the small social cues that say this person is reliable. It takes years, and it feels human, because it is.
Online, trust is something different. It's an algorithm. Not metaphorically, actually.
Every system that decides who gets seen, cited, recommended and believed is solving the same problem: how do you estimate something as fuzzy as trustworthiness when you can't shake someone's hand? The answer, everywhere I have looked, is the same three things.
Three clocks above the pillars
Attention works in seconds, proof works in minutes, reputation works in years, and the three pillars below are what you can work on at each speed. Brand is the proof you make. Trust Signal is that proof made legible to a machine. Reputation is the one you cannot issue yourself. The three clocks is the philosophy; this page is how machines read it.
A machine can read two of the three clocks. Proof can be checked, because a specification, a tally or a structured credential is a fact a crawler can verify. Conferred reputation can be counted, because a link is a vote somebody else cast and a review is a sentence somebody else wrote. Attention cannot be read at all, because nothing in a search engine's quality framework rewards the seconds clock, and that single fact explains why going viral moves rankings not at all, and why the businesses that rank are so often quiet ones with twenty years of kept promises and a website somebody finally made legible. In Google's own terms, Experience and Expertise are proof, Authoritativeness is conferred, and Trustworthiness is the pattern over time.
The Medieval Version
England, around 1300. A silversmith couldn't sell his work without a hallmark, a stamp pressed into the metal that could be traced back through a registry to the maker. If the piece turned out to be under weight, the blame had somewhere to land. Accountability was locked into the object itself.
That one stamp did three jobs at once.
It was a claim. This maker says this is their work, their name on the product, their reputation staked on its quality. That is what we now call Brand: what you say about yourself.
It was verifiable. The mark could be checked against a registry and matched to known makers, which meant even an apprentice with the ledger open could tell you whether a piece was genuine. That is what we now call Trust Signal: evidence a system can read, whether the system is Google or a boy with a book of marks.
It was endorsed. The guild stood behind the mark, and the guild had a reputation of its own, built over decades. That reputation took the hit whenever a hallmarked piece turned out to be a fake. That is what we now call Reputation: what others say about you.
Three channels, one outcome. You could trust a stranger because the system said so. Nobody had to know the silversmith personally, and nobody had to test every piece before buying it. The system was verifying credibility on everyone's behalf, which is what let strangers transact with confidence at all.
Seven hundred years later the architecture is identical. Google can't see the work you've done, can't shake your hand, can't read your character the way a medieval merchant could by asking the neighbours, so it reads three things instead.
What Google Actually Measures
Around 2023 Google formalised something it had been doing for years and called it E-E-A-T: Experience, Expertise, Authoritativeness, Trustworthiness.
The SEO industry treated it like a content instruction, make sure your content demonstrates expertise, tick the box, move on. I read it differently. E-E-A-T was never an instruction, it was a description of how the algorithm reads credibility. And not only Google's algorithm. LinkedIn works on the same pattern, so does YouTube, so do the recommendation engines, the ad platforms and the news feeds. Every system that has to guess at credibility without meeting you in person solves it the same way.
It maps to three pillars.
Brand is what you say about yourself: the website, the content, the positioning, the narrative you own and control. In E-E-A-T terms this is Experience and Expertise, the claim about what you know and what you have done. When a financial adviser writes about mortgage strategy on the company blog, that is Brand, their expertise made visible in their own words. You control this signal entirely, which is both its power and its limitation.
Reputation is what others say about you: the reviews, the backlinks, the mentions in places that matter, the signal you can influence but never fully control. In E-E-A-T terms this is Authoritativeness, the external validation that confirms the claim, and it carries weight precisely because it cannot be bought from yourself. I sold BookPrint, my book-printing business, about six years ago, and since then it has never spent a dollar on advertising. The new owners signed up to Trustpilot, the good reviews stacked up, and I believe that is the only marketing they have done. What other people said turned out to be worth more than anything we ever said about ourselves.
Trust Signal is what machines can read about you: schema markup, structured data, the technical layer that makes credibility discoverable to the systems doing the ranking and the recommending. In E-E-A-T terms this is the infrastructure that makes Trustworthiness verifiable. An adviser might carry twenty years of credentials, but until those credentials exist as structured data, Google can't tell them apart from a first-year graduate with a nice website. Trust Signal is the plumbing. Plumbing is invisible, which is exactly why it gets skipped.
Three pillars, each one measurable, each one buildable. When all three align, something compounds. When they don't, something leaks.
Trust Interpreter, Not Truth Engine
The Trust Algorithm doesn't determine truth. It approximates what a discerning human would consider trustworthy, at internet scale.
Think about buying from a business you've never heard of, a coffee subscription, a piece of software, shoes from a brand that arrived through an ad. I do what everyone does: read a few reviews, check for a physical address, glance at the checkout page for the security padlock, sometimes search the company name with the word scam attached, and form a judgment inside thirty seconds from whatever the screen offers. That whole compressed assessment is what the algorithm automates for billions of searches a day, and it reads the same three channels a person does.
Brand: does this business claim expertise that seems credible, is the positioning specific or generic, does the content reveal genuine knowledge or could anyone have written it?
Reputation: do other credible sources confirm the claim, are there backlinks from places that matter, do the reviews read like real people?
Trust Signal: can the claim be checked against structured, machine-readable evidence, is the data clean, can a machine parse who this person is and whether the credentials match the story?
The algorithm isn't judging whether you're a good person. It's reading whether your credibility is coherent across the three channels. When the story you tell, the story others tell, and the story machines can read are the same story, the output compounds. When they diverge, the break shows up as a feeling: something seems off, the visitor can't name it, and they quietly choose someone else.
There is no dramatic failure, just a slow leak, conversion at 2% when it could be 5%, bounce a little high, leads that arrive but never close, metrics that look mediocre rather than catastrophic. Mediocre never gets diagnosed as a trust problem, so the website gets redesigned, the prices get lowered, the ad spend goes up. None of it helps, because the underlying issue is coherence and everyone is busy treating symptoms.
How the Pillars Interact
Most businesses treat the three pillars as separate projects: Brand belongs to marketing, Reputation belongs to PR, Trust Signal belongs to a developer somewhere. Three workstreams, no coordination, and no sense that the pillars amplify or undermine each other.
Strong Brand with weak Reputation is theatre. A beautiful website making impressive claims that nobody else confirms reads to the algorithm as an unverified claim, and to a human as an unvalidated promise. The gap between what you say and what others say produces a scepticism that never becomes conscious, the kind that closes the tab without quite knowing why.
Strong Reputation with weak Trust Signal is invisible excellence. Trusted by everyone who has worked with you, recommended by colleagues, and unreadable to machines, because the credentials were never structured and the expertise was never made discoverable. This is the pattern I have seen most in professional services: decades of genuine credibility built face to face, and a digital presence that is an afterthought. A website built five years ago and never touched since. No schema, no structured data, nothing that tells Google this is a qualified professional whose credentials check out.
Strong Trust Signal with weak Brand is a well-formatted ghost. Machines can read the site perfectly, and humans arrive to find nothing specific, no opinion, no positioning, no evidence that anyone here understands their problem. The technical infrastructure is sound, and there is nothing worth discovering inside it.
You need all three, though not necessarily in equal measure, and where to invest depends on which pillar is weakest, because the weakest pillar is always the constraint. The other two can be excellent and it will not matter.
Two Businesses, Same Problem
The clearest way I know to show the pillars failing differently is to walk through two contrasting constructions. Treat them as worked examples, one from each end of the visibility problem, rather than as case studies.
Consider a deep-tech manufacturer. Superconducting magnets, say, one of the smallest niches there is, where the search volume across every relevant keyword might be a hundred qualified searches a month. A single conversion can be worth hundreds of thousands of dollars, because nobody googles superconducting magnets idly. Suppose the Brand pillar is weak: a generic website, no published depth, nothing online that demonstrates the knowledge sitting inside the company. The people in it understand the field as well as anyone; the internet has no way to know. Suppose Reputation is weaker still. No citations from physics publications, no industry mentions, no external voice saying these people are the reference. And suppose Trust Signal is functional but bare, schema missing, structured data absent.
The gap between what a company like that knows and what the internet can see of it is enormous, and gaps like that get filled, usually by a competitor.
The fix is not advertising, it is publishing the depth. Dense technical guides, honest comparisons, application notes, all of it linked to peer-reviewed research, with mentions earned from publications that carry credibility of their own. Positioned not as a company that sells magnets but as the reference for the technology. In a market of a hundred qualified searches a month, a citation from a serious physics journal doesn't just improve a ranking, it makes you the obvious choice. There is no noise to compete against, and the signal is clean.
Which pillar failed first? Reputation. Without external validation, the strongest Brand and the cleanest Trust Signal can't carry a buyer who is about to spend six figures with a company they have never met. Reputation is also the slowest pillar to build and the most permanent once built. Nobody can fake being recognised by their peers, and nobody can hurry the relationships that produce genuine citations.
I have lived a small version of that mechanism. At BookPrint we once quoted a corporate history book for a customer I had never met, and her existing printer, the company that did all her other work, had the same machine I had plus a whole lot more, literally the same machine, and gave her a cheaper quote. She came back, told me her printer was cheaper, and said she wanted to go with us anyway, because we were the book specialist. Everything she could find of us said books and only books, and that coherence beat a bigger company with a better price.
Now consider the opposite failure. Suppose an independent mortgage advisory firm with a strong Reputation, trusted advisers, established relationships, real credentials earned over years, and a decent Brand, coherent positioning, known in its industry. But no Trust Signal: schema wrong, structured data absent. When someone searches for an independent mortgage adviser, these advisers do not appear, not because they lack credibility but because machines cannot read their credibility.
The fix here is plumbing. Mark up every adviser properly, education, credentials, specialisations, locations, cleanly enough that Google understands not just what the firm says but what the advisers verifiably are. Nothing about the expertise changes, and nothing about the reputation changes. What changes is that the machine can finally read both, and in a market where machines deliver the first few results and humans choose among them, unreadable means invisible.
Which pillar failed first here? Trust Signal, and I would argue it is the most common failure in New Zealand. I have done marketing for NZ mortgage advisers, and the pattern sits right across our professional services. A small, relationship-driven market with a low tolerance for bullshit, full of firms whose credibility was built over decades of face-to-face work and whose digital presence is an afterthought. They are not short of trust. They just have not translated their trust into a format machines can read, so they are invisible by default, opaque by accident.
And it is not only Google doing the reading. LinkedIn reads structured signals, YouTube weights creator credibility, ad networks decide who is credible enough to advertise. Review platforms index structured data, industry databases cross-reference credentials. Leave those signals unclaimed and you are invisible across all of them at once.
Two constructions, opposite failures, one architecture. In both, something genuinely valuable is invisible online, and the work is to find the broken pillar and fix it. A repaired pillar gives the other two something to amplify, while a broken one leaves them working hard for nothing.
When All Three Pillars Collapse
The most instructive public case is the one where all three pillars were faked at once.
FTX.
Brand: Sam Bankman-Fried was everywhere, magazine covers, congressional testimony, the casual-genius narrative, the responsible crypto billionaire cleaning up his own industry. Young founder, billionaire by thirty, worried about existential risk, giving the money away. The story was completely coherent, and he told it constantly.
Reputation: Sequoia Capital invested, and that is not just a cheque, it is institutional validation, the kind that makes other institutions follow. Celebrity endorsements lined up, and the admiring profiles ran in major publications. Every channel of external validation said the same thing: this is the most legitimate crypto company in the world.
Trust Signal: a corporate structure that looked exactly the way a financial institution should look, regulatory registrations, banking relationships, a polished institutional surface that read as official to every system checking it.
Three pillars, perfectly aligned, all fake.
Then it collapsed, all three at once, not sequentially and not gradually. The Brand evaporated when the fraud emerged, and years of narrative-building vaporised in weeks. The Reputation reversed: the same publications that had praised him became his prosecutors, Sequoia wrote its stake down to zero, and the endorsements turned into public distance-taking. The Trust Signal was exposed as surface the moment anyone audited what sat underneath the institutional look.
You can only fake all three pillars at once for so long. The collapse was total because nothing real sat underneath. No genuine operation to sustain the Brand once the fraud emerged, no independent scrutiny that could have caught the problems earlier, no machine-readable claim grounded in where the money actually was.
The question FTX forces on the rest of us is a useful one. What happens if someone looks closely at the pillar where you are weakest? What if a journalist, a competitor or a regulator decides to audit exactly where your credibility is thinnest?
The Scale Question
Trust-Promise Pairs operate at micro level: one person, one promise, one delivery. Promise a five-minute read and deliver a five-minute read, and a small trust transaction has occurred, no money changing hands. The Trust Algorithm operates at macro level, the system-wide read of your credibility across all the evidence available. They are the same model at different scales.
The connection between the scales is direct. Each kept promise generates a signal, a review, a return visit, a share, a backlink from someone who found the thing genuinely useful, and those micro-signals accumulate into macro-credibility. The trust account fills one small transaction at a time, and the algorithm reads the balance.
I learned the durable end of this at nineteen, running a lawn-mowing round and advertising it with flyers, 10,000 at a time, black and white because the colour ones tested worse. Six or seven years after I stopped mowing lawns, people were still ringing the number from a flyer stuck to a fridge. A plain flyer that says exactly what you do is a very small trust machine, and that one kept working from the fridge door for most of a decade. The algorithmic version is the same shape. It is just written where machines can read it.
Understand how the macro level reads and you can build the micro level for it, shaping the content, the customer service, the follow-up emails and the technical foundation so they reinforce one another. Every touchpoint becomes a small deposit that compounds into algorithmic credibility. Skimp on any pillar and the weakness surfaces somewhere unexpected. A beautifully designed website with no external validation makes people hesitate at the exact moment of conversion. Years of trusted relationships with no technical infrastructure leave you invisible to the systems that deliver new clients. Perfect schema wrapped around no real expertise makes you rankable but unconvertible, found and abandoned in the same minute.
Most people optimise one pillar, usually Brand, because it is creative and controllable, and hope the others follow. They don't.
The Imbalance Problem
Most businesses are wildly unbalanced across the three pillars.
Brand gets all the love. The beautiful website, the polished messaging, the content strategy that tells the story the business wants told about itself. Marketers gravitate here because the work is visible: the designs can be seen and approved, the output can be pointed at, and progress feels like it happened. Most marketing budgets go here, and most agency work lives here.
Reputation gets neglected, because genuine third-party validation is slow, uncertain and unglamorous, and because nobody can schedule it. Earned mentions and credible backlinks refuse to fit inside a project plan, no one can guarantee them by Q3, and so the work gets deprioritised, every quarter, indefinitely.
Trust Signal is usually unknown territory. Schema markup, structured data, the technical foundations that make everything else discoverable. Most business owners do not know the layer exists, and most marketing teams lack the technical knowledge to build it, so it gets filed under someone else's problem and stays there. The competitors who understand it stay visible.
The result is predictable: impressive brands nobody can find, genuine reputations machines cannot read, technically perfect websites wrapped around an empty core. All of them spending money, all of them working hard, none of them compounding, because the system is unbalanced.
Your weakest pillar is where competitors eat you, not because they are better, but because they are more coherent. The one who found their weakest pillar and fixed it first will outpace the one with the bigger budget pouring money into the wrong dimension.
Fixing the weakest pillar is usually the cheapest intervention with the highest return. Trust Signal work is mostly a one-off technical investment, Reputation building is slow but does not need a big budget, and Brand sharpening is editorial work rather than a redesign. The fix is rarely expensive. It is just invisible until you know to look for it.
Find your weakest pillar: /diagnostic/
Why This Matters Now
We are in a strange moment, where exposure costs nothing and most businesses still treat trust as an accident.
AI has made noise essentially free to produce. Fake reviews are commoditised, authority can be manufactured for a season, and anyone with a chatbot subscription can produce content that looks expert at a glance. The theatre has never been cheaper to stage. At the same time, every algorithm update, every audit tool that gets easier to run, every forum thread that unpicks a fake-review farm narrows the gap between theatre and the real thing. The cost of maintaining the act keeps climbing while the value of genuine signals compounds.
Ten years ago a website and some basic content could rank, because the barrier to entry was low. Now every niche is crowded, every vertical has someone with funding and decent design, and yet most of them are still under-invested in the actual Trust Algorithm. Plenty of money on Brand, some accidental Reputation if the product happens to be good, and Trust Signal ignored completely because it is not glamorous and it requires technical knowledge.
I should be clear about why I keep arguing for the honest version, because it is not piety. I do not need honesty to be virtuous, I need it to work, and it does: any departure from truth eventually forces the customer to reclassify you, as ignorant, as incompetent, or as something worse, and each of those reclassifications ends the relationship in its own way. If a tactic only works while the customer does not know about it, that is not marketing, it is manipulation, and it is carrying its own collapse around inside it.
The businesses that survive this transition are the ones that understand trust is structural, not a vibe and not a brand refresh, a system with measurable inputs and measurable outputs. Brand plus Reputation plus Trust Signal go in; visibility, credibility and conversion come out. When the three inputs cohere across every channel, the output compounds until, over months and years, the position becomes genuinely hard to displace. Not through gaming anything, but because you are the signal: the same story told by you, confirmed by others, readable by machines.
The Trust Algorithm doesn't reward the loudest voice. It rewards the most coherent signal.
Faking it works briefly. Real trust work is slower, and the slowness is the point, because it has momentum: you do not wake up wondering when the collapse comes. Every fake that gets caught raises the value of whatever is real. Every company that folds under scrutiny makes genuine authority worth a little more to the companies still standing.
Everything on this page renders from a seed file I author and maintain, the same seed that produces five sites and roughly eighty pages, none of which I edit by hand. Marketing Curious: Working the Noise is where the thinking runs at full length. This page is a rendering. The seed is the source. The book is the story of building it.
Where to Go Next
The three clocks is the philosophy above this page, and attention is the clock this algorithm cannot read.
Understand each pillar: Brand | Reputation | Trust Signal
Find your weakness: The Diagnostic
See the dark side: Authority Theatre
Build from the bottom up: Trust-Promise Pairs