Methodology
How do we measure? We show our work.
This product's entire promise rests on the measurement being real. So we don't hide what we measure, or how.
We ask real questions
We ask each engine, live, the kinds of questions your customers would ("Who's the best Y in city X?"). The questions come from fixed templates, filled in with your category, city, and district.
Web search is on
All 3 engines we query live run with web search / live-source access enabled. The goal is to measure the live answer an engine gives from what it reads on the internet today — not what it memorized.
Each answer is sorted into three states
Every answer is parsed by a separate model: Recommended (you appear as a recommendation), Known (it knows you but doesn't recommend you), or Not mentioned (you don't come up at all). Name matching is meaning-based — "Özkan Dental" and "Özkan Oral & Dental Health" count as the same business.
Your score is calculated
Engine results, your ranking, and how readable your site is to AI engines are distilled into a single score from 0 to 100. It is re-measured every Monday; your weekly change, 4-week average, and 12-week trend are all stored.
Honesty note
We query the engines' official APIs. These answers may not be 100% identical to a chat in the same engine's consumer app — models, versions, and search behavior can differ slightly between surfaces. The answer you see on the results screen is real and live, not a simulation; but it is a representative measurement, not a single absolute truth. That is exactly what makes progress meaningful: because we measure the same way every week, week-over-week comparisons stay consistent.