Web analysis often treats links as if they were a direct proxy for brand visibility.
The Brand Mention research shows why that assumption is too simple.
Gold Dataset v1 separates three different structures:
416 Explicit Brand Mention — Linked
328 Unlinked Brand Mentions
378 Linked Entity References
Together, the explicit Brand Mention layer contains:
744 explicit Brand Mentions
Among those explicit mentions:
44.1% are unlinked
That means visible brand naming and hyperlink presence overlap.
But they are not the same thing.
The Difference Between Being Linked and Being Named
The full article continues below in its original wording.
Being linked is a structural property
Read section →02Insight sectionBeing named is a visible-text property
Read section →03Insight sectionThe 328 unlinked Brand Mentions change the picture
Read section →04Insight sectionThe 378 Linked Entity References show the opposite problem
Read section →05Insight sectionThe same page can contain both structures
Read section →06Insight sectionWhy visible naming deserves its own measurement
Read section →Being linked is a structural property
A link tells us that one piece of content points somewhere else.
That destination may be:
a company page,
a product page,
a service page,
or another entity.
The link creates a connection.
But the visible anchor does not always contain the brand name.
For example:
"See this model"
can link to a Volvo-related destination.
The entity is linked.
The brand is not visibly named in the same way.
That is why the taxonomy includes:
Linked Entity Reference
as a separate class.
Being named is a visible-text property
A Brand Mention occurs when the brand name is visible in the analysed text.
For example:
"Volvo is worth considering."
The brand is named directly.
That mention can be:
linked
or
unlinked.
The important point is that visible naming exists independently of hyperlink structure.
The 328 unlinked Brand Mentions change the picture
In Gold Dataset v1:
328 explicit Brand Mentions are unlinked
That equals:
44.1% of the 744 explicit Brand Mentions
A link-only visibility report would miss those records.
The brand is clearly present in the text.
The reader can see the name.
But there is no hyperlink attached to the mention.
That makes unlinked Brand Mentions an important part of external brand visibility.
The 378 Linked Entity References show the opposite problem
The other error happens when every link is treated as an explicit Brand Mention.
Gold Dataset v1 contains:
378 Linked Entity References
These records contain a relevant entity link without equivalent visible brand naming.
If every outbound link were counted as an explicit Brand Mention, the visible naming layer would be inflated.
This means link-based reporting can fail in both directions.
It can:
miss unlinked mentions
and
overcount descriptive or entity references.
The same page can contain both structures
A discussion can contain:
a visible unlinked Brand Mention,
a linked explicit Brand Mention,
and a Linked Entity Reference
inside the same page.
Those structures are not mutually exclusive at page level.
The important distinction is the classification of the individual evidence unit.
This is why page counts and Brand Mention counts should not be treated as interchangeable.
Why visible naming deserves its own measurement
When a brand name appears directly in the text, the reader encounters the entity explicitly.
The brand is not hidden behind:
generic anchor text,
a product label,
or descriptive wording.
That makes visible naming a distinct structural signal.
The research does not claim that visible naming is always stronger than linking.
It claims only that the two should be measured separately.
Why linking still matters
Separating naming from linking does not make links unimportant.
A hyperlink can:
identify a destination,
connect a mention to a specific product or service,
provide entity resolution,
or add navigational context.
The problem begins only when link presence is treated as equivalent to visible naming.
Both structures can be informative.
They answer different questions.
A simple three-case comparison
Case 1 — Named, not linked
"Volvo is worth considering."
Visible brand name.
No hyperlink.
Classification:
Unlinked Brand Mention
Case 2 — Named and linked
"Volvo is worth considering."
The word Volvo is linked.
Visible brand name plus hyperlink.
Classification:
Explicit Brand Mention — Linked
Case 3 — Linked, not explicitly named
"See this model."
The phrase links to a Volvo-related destination.
The brand name is not visibly present.
Classification:
Linked Entity Reference
All three structures can connect the discussion with the same brand.
They are not equivalent evidence.
Why this matters for measurement
A good measurement system should be able to answer at least three separate questions:
Was the brand visibly named?
Was the visible mention linked?
Was an entity linked without equivalent visible naming?
If those questions are collapsed into one metric, the report becomes easier to read but harder to interpret.
The taxonomy exists to preserve that structure.
Why this matters for reputation analysis
External brand visibility is not limited to clickable references.
A brand can accumulate a large amount of visible textual presence without receiving links.
That matters because brand perception develops through language as well as navigation.
A person reading:
"Volvo is reliable"
has encountered the brand claim whether or not the word is clickable.
This does not prove that an unlinked mention has greater influence.
It shows that visibility exists outside hyperlink structure.
Why this matters for distributed brand assets
A Brand Mention can exist on a publisher-owned page the brand does not control.
If the mention is unlinked, the brand may still receive:
visible naming,
context,
association with products or problems,
and presence in a decision conversation.
That is one reason external Brand Mentions can behave like distributed brand assets in a structural sense.
The asset is the externally hosted brand context.
The brand does not necessarily own the page, URL or surrounding editorial environment.
Why this matters for AI Search research
AI systems can encounter both:
explicit brand names
and
linked entities.
The current research does not establish that one structure is more influential on AI systems than the other.
It creates a measurable distinction that can be tested.
Future research can ask:
Do linked explicit mentions behave differently from unlinked explicit mentions?
Do Linked Entity References contribute differently from visible naming?
Does repeated unlinked naming appear in AI-generated brand descriptions?
Those remain empirical questions.
The current study does not assume the answer.
Link equity and Brand Mention evidence are different concepts
A hyperlink can have technical properties relevant to web systems.
A Brand Mention can have semantic properties relevant to visible text and context.
Those are not interchangeable concepts.
The current research focuses on Brand Mention evidence.
It does not use hyperlink presence as a shortcut for semantic visibility.
This distinction becomes especially important in AI Search research, where the question is not only:
"what points to the brand?"
but also:
"what does the surrounding text say about the brand?"
Why one metric cannot replace the other
If we report only links, we lose unlinked visible naming.
If we report only visible naming, we lose entity-linking structure.
The strongest reporting model keeps both.
That allows analysts to distinguish:
brand presence,
link structure,
and entity connectivity.
The result is more precise than a single blended count.
What this insight does not claim
This insight does not claim that unlinked mentions are better than linked mentions.
It does not claim that linked mentions are stronger than unlinked mentions.
It does not claim that Linked Entity References are weaker.
It does not claim that hyperlink presence causes greater AI visibility.
It does not claim that unlinked mentions are more organic.
And it does not claim that the 44.1% unlinked share applies to the entire Full Research Inventory.
The conclusion is narrower:
being linked and being visibly named are related but distinct forms of brand presence, and they should be measured separately.
The practical takeaway
A link answers:
Is there a navigable connection to an entity?
A Brand Mention answers:
Is the brand visibly named in the text?
Those questions overlap.
They are not identical.
That difference is why link counts cannot replace Brand Mention counts.