Building Authority Across Multiple Brands
How to stop AI treating your other websites as unrelated, unproven strangers to each other
The short version
Everything in this series so far has assumed one business, one site. If you run more than one — a company operating two or three separate brands, each with its own domain and its own audience — the rules don't change, but a new failure mode opens up: you can do everything right on each individual site and still end up weaker than if you'd built one strong site, because the sites work against each other instead of for each other.
The fix isn't "pick one brand and focus." It's understanding which parts of your authority belong to the company that owns everything, and which parts have to be built separately for each site — and not confusing the two. Get this right and every new brand you launch makes the others easier to trust. Get it wrong and every new brand dilutes what you've already built.
The rule this guide is built on: build the entity first, the content second. Company once, site each.
Why this is a different problem, not just "do Guide 1 three times"
Authority, as the rest of this series has covered it, is a web of independent, machine-readable references that all agree with each other. An AI system trusts you when your own site, other people's sites, business registries, and review platforms all describe the same thing the same way.
With one business, that's one identity to keep consistent. With three brands under one company, you now have three identities that need to stay individually clear and need to visibly connect back to the same trustworthy source behind all of them. Miss that second part and an AI system has no way to know your insurance-comparison site and your personal-finance site are backed by the same accountable business — it just sees two unrelated, individually thinner entities, each with less evidence behind it than the company as a whole actually has.
The three layers (formal label: entity layers)
Every business trying to be found and cited by AI systems is really building three separate entities, whether it thinks of it that way or not. The question this guide answers is which of the three you build once, centrally, and which you have to build again for every brand.
| Layer | What it actually is | Built once, or per site? |
|---|---|---|
| Person — the author | A real, identifiable individual who can be checked and cross-referenced, not just a byline | Depends on your structure — see below |
| Company — the entity behind it | The legal business, verifiable against outside registries other systems cross-check | Once. One company, many sites |
| Website — the page itself | The on-page and technical signal stack for that specific site | Per site. Every brand needs its own |
The company layer is the one businesses running multiple brands most often get wrong — not by neglecting it, but by rebuilding it inconsistently, half-connecting it to some sites and forgetting others. More on that below.
Layer one: the company (build once, reference everywhere)
(Formal label: entity consolidation)
If one legal entity owns several brands, that entity should exist, provably, in exactly one place — and every site it owns should point back to that same place the same way. This is the layer where "once" is not just efficient, it's correct: registering the same company twice, slightly differently, under two different brand names is worse than registering it once, because it gives outside systems two competing, half-matching records to reconcile instead of one clean one.
What belongs at company level:
| Mechanism | What it does |
|---|---|
| Business registry listing (e.g. Companies House or your jurisdiction's equivalent) | The base legal fact that everything else gets checked against |
| A structured open business database entry (e.g. OpenCorporates-type listing) | Machine-readable confirmation of the same legal fact |
| A knowledge-base entity item (e.g. a Wikidata-type item) | A neutral, third-party-editable record AI systems treat as strong corroboration |
| A business-intelligence profile (e.g. a Crunchbase-type listing) | Another independent point of confirmation, useful mainly for the Google Knowledge Graph side of things rather than direct AI citation |
| Consistent identifiers — company number, VAT number, any relevant regulatory ID | The thread that ties every other record back to the same entity, even when the brand name differs |
| Review/business-profile platforms | Where genuine, moderated third-party opinion about the business lives |
Every site the company operates should reference this same set of identifiers and the same company name, in the same form, in its schema. That's the entire mechanism: not that each site does its own version of this work, but that each site links to the one version that already exists.
Layer two: the website (build separately, every time)
(Formal label: on-page signal stack)
Each brand's site is a distinct thing an AI system has to independently decide it trusts. It doesn't inherit trust automatically just because its owner is well established elsewhere — it has to earn its own on-page and technical foundation, in full, the same way a standalone business would.
| Mechanism | What it does |
|---|---|
A fully linked schema graph (Organisation, Person, WebSite, WebPage sharing consistent @id references) |
Ties the page's content to the people and company behind it in a form machines can follow |
| Canonical URLs | Removes ambiguity about which version of a page is the real one |
sameAs links out to the company's and author's other verified profiles |
The connective tissue back to layer one and layer three |
llms.txt, a working robots.txt, IndexNow submission |
Baseline AI-crawler accessibility — covered in earlier guides in this series |
| Sitemap and, where the site publishes regularly, RSS | Discovery and freshness signals |
| Standard trust pages — About, Contact, Privacy, Terms | Table stakes; their absence is itself a negative signal |
| A glossary or defined-terms structure where the site's subject matter warrants it | Helps AI systems extract precise, citable definitions rather than paraphrasing loosely |
| Review or rating markup, only where genuine reviews exist | Corroboration — never fabricate this layer |
None of this is brand-specific creative work — it's the same checklist Guide 1 through 8 already walked through. The point for a multi-brand operator is just that it has to be done in full, separately, for every site. There's no shortcut where a strong company entity substitutes for a thin website.
Layer three: the person (the layer you have a real choice about)
(Formal label: author entity)
This is the layer earlier guides in this series treated as settled — one named, credentialed author, consistently represented. Running multiple brands is where that assumption breaks, because you now have a genuine structural choice to make, and it's the question this guide exists to help you answer.
The three options
Option A — one strong personal entity, fronting every brand. A single named individual is the visible author/editor across all the sites. Every brand's About page, byline, and schema points to the same person, the same credentials, the same LinkedIn-type profile.
- Strength: concentrates all your author-authority building into one entity instead of splitting effort three ways. Faster to build a resolvable, well-corroborated person record because you're only building one.
- Weakness: if the brands cover unrelated subject areas, one person claiming expert authority across all of them can look thin or opportunistic — both to a human reader and to a system evaluating subject-matter authority. It also means every brand's credibility rests on one person's.
Option B — a different named author per brand. Each site has its own editor or author, genuinely specialised in that site's subject.
- Strength: each author's credentials map cleanly onto their site's actual subject matter — cleaner subject authority per brand, and no brand is a stretch for its named author to speak on.
- Weakness: you're now building three separate person entities from nothing, each needing its own profile, credentials, and cross-references. More work, and none of them individually gets the benefit of the others' accumulated trust.
Option C — hybrid: the company is the visible byline, with named contributors where it earns credibility. The company itself is the consistent author of most content; specific named individuals are credited only where a real credential adds something (a qualified specialist writing in their actual area, a founder commenting on company-specific matters).
- Strength: avoids overclaiming personal expertise the business doesn't have, while still using named authority exactly where it's genuine. Scales cleanly as you add brands, since you're not required to manufacture a new personal entity for each one.
- Weakness: a company byline alone is a weaker trust signal than a real named author — Guide 1 covered why AI systems weight credentialed individuals more heavily than institutional voices. Used everywhere, this option quietly gives up one of the stronger trust signals available.
What actually decides between them
There's no universally correct answer here — it genuinely depends on the business:
| If your situation is… | Lean toward… |
|---|---|
| One person with real, provable expertise that genuinely spans all your brands' subject areas | Option A |
| Brands in genuinely distinct fields, each better served by a specialist who can credibly speak to that one subject | Option B |
| A small team, limited time to build person entities from scratch, but a few individuals with real, narrow credentials worth surfacing selectively | Option C |
| Early-stage, brands still finding their footing, no individual yet has enough of a track record to be worth building into a public entity | Option C, provisionally, revisited later |
Whichever you choose, apply it consistently within a brand. Mixing signals on a single site — sometimes the company speaks, sometimes an uncredited "team," sometimes a named author who then disappears from later pages — recreates the exact inconsistency problem Guide 1 warned against, just at a smaller scale.
Which platforms are inherently personal vs inherently business
This is a decision each mechanism forces on you, whichever option you pick overall:
- Inherently personal, follows the individual regardless of brand: a personal profile on a professional network, an individual researcher identifier (e.g. ORCID-type), any platform built around individual credentialing or reputation, personal participation in public forums or Q&A communities.
- Inherently the business, not a person: the company's registry listing, its business-profile and review platforms, its structured business-database entries, any operator or brand-level social accounts.
- Genuinely both: a professional network often supports both a personal profile and a company page — and does its own job better when the two reference each other, rather than when only one is maintained.
A useful general caution on the personal side: entity records on open, third-party-editable knowledge bases (the kind AI systems treat as strong independent corroboration) typically require the subject to already be independently notable — referenced by sources the individual doesn't control — before the record is accepted. Creating one for a person before any independent coverage of them exists tends to get it rejected or removed rather than accepted early. Company entity records don't usually carry the same bar. This is a reason to sequence person-layer work rather than a reason to skip it.
Worked example: one company, two brands
To make the three layers concrete, here's a fictional company — invented for this guide, no relation to any real business — running two separate consumer sites.
The company: Thornfield Financial Ltd, a UK financial-content company. Brand one: PennyWell — a personal budgeting and saving-money site. Brand two: CoverScope — an insurance-comparison site.
Different subject matter, different audiences, same legal owner.
| Built at company level (once) | Built per site | |
|---|---|---|
| Company entity | Thornfield Financial Ltd registered on the business registry, the open business database, and the knowledge-base item; company number and VAT number consistent everywhere | PennyWell and CoverScope both cite "operated by Thornfield Financial Ltd" in schema and footer, with matching identifiers — neither re-registers the company under its own name |
| Website stack | — | PennyWell and CoverScope each have their own full schema graph, canonical URLs, llms.txt, sitemap, About/Contact/Privacy/Terms, and — because both genuinely collect them — their own review markup |
| Author | — | Thornfield chooses Option B here: PennyWell is fronted by a named personal-finance editor with real budgeting/consumer-finance credentials; CoverScope is fronted by a different named editor with an insurance background, because the subjects don't overlap enough for one person to credibly claim both |
The failure mode this guide exists to prevent: Thornfield registers the company on the knowledge-base item once, links it from PennyWell's schema, and never gets round to adding the same reference to CoverScope. To an AI system, PennyWell now looks backed by an established, verifiable company — and CoverScope looks like an unrelated, unbacked site that happens to share a footer credit. The company's own real credibility never reaches the second brand, not because it doesn't apply, but because nobody finished connecting it.
That's the pattern worth watching for in general: a business builds the company-level entity properly once, applies it fully to whichever brand happened to launch first or matters most at the time, and never goes back to wire the others in. It isn't usually deliberate — it's what happens when "add the company entity" is treated as a page one-off rather than a checklist item for every site the company ever launches.
Other common versions of the same mistake, worth checking for on your own brands: reusing one named author across brands without updating their bio or credentials to match each site's actual subject (so the "same person" appears to be an unrelated kind of expert on each one); giving different brands genuinely different company descriptions instead of the same one reworded for each audience; and letting one brand collect real third-party reviews while another has none, then treating the whole company as review-corroborated when only a third of it actually is.
The one-line test
Before adding a new brand, or auditing existing ones: for any given fact about the company — its legal name, its number, who owns it, who's accountable for what it publishes — can an outside system find that fact stated the same way from more than one independent site or source, across every brand you operate, not just your best one? If the answer is "only on one of them," that's the gap to close before building anything new.
How the nine guides fit together
This is the last guide in the series, so here's the shape of the whole thing.
Guide 1 established the core distinction the rest of the series rests on: AI search is a different system from traditional search, judging businesses on entity trust rather than link authority, and named that trust breaks down into five checkable signals. The guides in between took that model and turned it into practice — building the technical foundation an AI system can actually read, establishing a real author entity behind the content, structuring and writing content that holds up to citation, and getting corroborated in the places outside your own site that make an AI system believe what you say about yourself.
This guide is where all of that gets stress-tested against a harder case: what happens when it's not one entity you're building, but several at once, sharing one owner. Everything from Guides 1–8 still applies in full to each individual site — nothing here replaces that work. What this guide adds is the layer above it: knowing that the company behind your brands is a single entity to build once and reference consistently, that each website still needs its own complete signal stack regardless, and that the author question has no default answer — it's a real choice between a single strong personal entity, brand-specific specialists, or a hybrid, made deliberately rather than by accident of who happened to write the first page.
If you've worked through the series in order, this is the point where the individual pieces stop being separate jobs and become one system: a company that's provably one thing, fronted consistently or deliberately-not, publishing through sites that are each, individually, as strong as if they were the only thing that company had ever built.