Module · Persona Intelligence
Questions built from keywords measure what a marketer already knows to ask. The questions that decide a purchase are the ones nobody typed into a keyword tool — is it trustworthy, X or Y, will I regret this.
The distinction
Every engine describes you when asked 'what is GoPersona' — you supplied the name. That is echo, not visibility, and it is not the question a buyer who has never heard of you asks.
A high score earned by measuring the easy thing
'Best CRM' is not a real question. A real buyer asks 'my team is five people and I want out of spreadsheets', and the answer names entirely different brands.
Correctly measured, but not your market
In Turkey the journey starts in social and AI is the verification step. A translated persona misses that order: it asks discovery questions when the real question is verification.
Right language, wrong behaviour
Our persona sets are built on the Censydiam motivation wheel, brand archetypes and 16 Turkey-specific personas. Which models we use and why is written out on /modeller.
How they are built
Every question ties back to a motivation — reassurance, belonging, control, discovery. The report shows which persona each question came from and where that persona loses you, so 'my visibility is low' becomes 'I am invisible to the cautious buyer'.
The persona sets are currently designed, not measured. No published corpus exists for how Turkish buyers actually phrase prompts; we are running our own experiment and will publish the result as it lands.
same brand · different persona
Illustrative breakdown. A single visibility percentage is the average of these four rows and describes none of them.
Per persona
Method
These sets are derived from a model, not mined from a real Turkish prompt corpus. Those are not the same thing and they do not get written the same way in a report.
The Censydiam motivation framework, brand archetypes, and 16 personas specific to the Turkish market. Which models we use — and which, like Human Design, we deliberately do not — is written out on /modeller.
Yes. On the enterprise plan you can define your own segments and fold them into the question set; once the panel is frozen the trend stays comparable.
No, and we say so. They are model-derived. A separate study is measuring real Turkish prompt behaviour, and the sets will be updated against it.
The free scan gives you the first persona breakdown and a first look at what the engines say about your brand.
Measure my brand free