The case studies exist for one reason:
to show what changes when the research methodology is applied to real data.
They are not client success stories.
They are not testimonials.
They do not try to prove that Brand Mentions caused a ranking, sale or AI recommendation.
Each case begins with a familiar analytical shortcut and then shows what happens when the source material is classified more carefully.
The useful question is not:
“What did FunkyMEDIA achieve for a client?”
It is:
“What do we understand differently after the data is separated into the right units?”
From Outbound Links to Explicit Brand Mentions
The case study shows why a link count and a Brand Mention count are not interchangeable, separating entity linking from explicit naming.
Repetition Changes the Meaning of Mention Volume
The case study shows that frequency and language diversity are different measurements and why repeated wording changes the interpretation of volume.
One Mention, Multiple Semantic Relationships
One passage can contain multiple structured relationships without pretending that several mentions occurred.
Question Before Recommendation
A brand can already be part of the decision process while a user is still asking, checking or comparing options.
Problem to Brand
Problem-centered conversations can be represented as structured evidence where the source text actually supports that path.
From Outbound Links to Explicit Brand Mentions
Gold Dataset v1 contains 794 outbound links.
A simple link audit could stop there and describe which destinations received links.
But once the visible text is classified, the picture changes.
The same dataset contains:
- 416 linked explicit Brand Mentions
- 328 unlinked Brand Mentions
- 378 Linked Entity References
The case study shows why a link count and a Brand Mention count are not interchangeable.
It demonstrates both directions of the problem:
some explicit Brand Mentions have no link,
and some entity-owned links do not visibly name the brand.
The result is a cleaner distinction between entity linking and explicit naming.
Repetition Changes the Meaning of Mention Volume
Gold Dataset v1 contains 744 explicit Brand Mentions.
If we stop at that number, every occurrence appears to contribute equally to the total.
Repetition analysis changes the interpretation.
Within those 744 explicit mentions:
- 63 were classified as exact formulaic mentions
- they formed 21 exact formulaic clusters
- 681 remained unique-like under the current method
The case study shows that frequency and language diversity are different measurements.
A total can remain unchanged while the information environment looks very different once repeated wording is identified.
This does not make repeated mentions worthless.
It makes the total easier to interpret.
One Mention, Multiple Semantic Relationships
A single Brand Mention can contain several different kinds of information.
One passage may name a company while also describing a product, problem, feature, use case, result and another brand.
The semantic calibration contains 536 relationship records and 276 unique Brand → Relation → Entity pairs.
This case study shows how one mention can be decomposed into several structured relationships without pretending that several mentions occurred.
The input is one piece of text.
The output is a set of separate semantic observations.
That distinction is the basis of Brand Context Graph.
Question Before Recommendation
Brand visibility does not begin only when somebody recommends a company.
The Gold Dataset calibration identified:
- 248 question / consideration signals
- 179 recommendation signals
The case study shows how a brand can already be part of the decision process while a user is still asking, checking or comparing options.
This changes the analytical focus.
Instead of measuring only final endorsement, the research can identify earlier conversational roles.
The case does not claim that a question equals purchase intent.
It shows that consideration and recommendation are different stages.
Problem to Brand
Some discussions begin with a brand.
Others begin with a problem.
The semantic calibration includes:
- 23 Problem relationship records
- 288 Use Case records
- 23 Result relationship records
- 64 Competitor / Co-Mention records
That makes it possible to study paths in which a user describes a need, encounters one or more brands and later discusses features, use cases or claimed results.
The case study is useful only where the source text actually supports that path.
It does not prove that the brand solved the problem.
It shows how a problem-centered conversation can be represented as structured evidence.
What these cases have in common
All five cases follow the same logic.
First, begin with a simple raw signal:
a link,
a mention count,
a sentence,
a question,
or a problem.
Then classify what is actually present in the source material.
Finally, compare the raw interpretation with the structured interpretation.
The point is not to create more metrics.
The point is to avoid forcing one metric to answer questions it cannot answer.
Why case studies are useful here
Definitions can be correct and still be difficult to understand.
A reader may accept that a Linked Entity Reference is different from an explicit Brand Mention without immediately seeing why that changes a report.
A reader may understand that repetition exists without seeing how it changes a total.
Case studies make those differences visible.
They show the transformation from:
raw record → classification → structured evidence → interpretation
That sequence is what turns methodology into something practical.
What these case studies do not prove
The cases do not prove that one mention type is universally better than another.
They do not prove that repetition is bad.
They do not prove that a semantic relationship is a verified external fact.
They do not prove that a question equals intent to purchase.
They do not prove that a Problem → Brand → Result path represents verified causality.
And they do not prove that Brand Mentions cause AI recommendations.
They demonstrate how the same source material can support more precise analysis once different structures are kept separate.
Where to start
If your interest is link analysis, begin with From Outbound Links to Explicit Brand Mentions.
If you are evaluating campaign volume, read Repetition Changes the Meaning of Mention Volume.
If you want to understand semantic context, read One Mention, Multiple Semantic Relationships.
If you are interested in the decision journey, read Question Before Recommendation.
If you want to see how need-based conversations can lead toward brands, read Problem to Brand.