# How is a measurement made? > A registered question set, buyer profiles and selected engines. Each question goes to each engine, each mention is identity-checked, the result has a range. Updated: 2026-09-29 A GeoMono measurement follows the same steps every time. The full method is on the [Method](/en/yontem) page; this page covers what a client needs to know. ## Question set and buyer profiles A measurement is made with a question set prepared for the brand's sector. The questions are written the way different buyer profiles would ask them (for example a buyer comparing prices, or a corporate purchaser). The question set is registered before the measurement and given a version number. The questions do not change in the middle of a measurement. ## How are the questions asked? Each question is asked separately to each selected engine, and the answer that comes back is stored as is. Questions on which the engines disagree are asked 3 more times to every engine that answered, and the majority decision is used; because the same question can get different answers at different times, no single answer is relied on. Questions that contain your name do not count towards the visibility rate. What those questions are for is explained in [Questions that name you and questions that do not](/en/yardim/sorular-ve-yanitlar/adiniz-gecen-ve-gecmeyen-sorular). ## Every mention is checked A name resembling yours appearing in an answer does not prove that you were named; it could be another company with the same name. So every mention is checked separately to confirm it belongs to your brand. If more than 20% of the mentions cannot be verified, no visibility rate is given and the report says "not measured". No rate is given either if there are not at least 5 different questions asked without your name: a rate calculated from a single question is not a rate. ## The result comes with a range A range is published next to the rate. The range says how stable the figure is under this question set; it is not a confidence interval for the brand's "true" visibility, because the questions were not chosen at random but written to a design. A low rate does not mean low demand or a low market share; it only says how often you were named in the answers to these questions. ## When are two measurements compared? Only measurements made with the same question set and the same version are compared. If the question set has changed, the report makes no comparison and says so plainly. The details are in [How to read the report summary](/en/yardim/raporlar-ve-paylasim/rapor-ozeti-nasil-okunur#period-comparison).