Module · Visibility Measurement

Are you invisible,
or just uncredited?

Both land in the report as the same low percentage. But one needs months of content investment and the other needs a single sprint. GoPersona tells you which.

Measure my brand freeNo signup · no card · 30 seconds

The distinction

Three different problems hiding inside one percentage

not found

The engine has never heard of you

Your name is absent and your site is not among the sources. This is a genuine coverage gap: publish, get cited, build the entity.

Months of content and authority work

not credited

Read, but never named

The engine answers from your pages and cites them as a source — and never says your brand in the sentence. The buyer never hears of you.

Naming and attribution on the page — one sprint

wrong entity

Someone else, under your name

Your name appears but the company being described is not you. A namesake is collecting your visibility, and every tool counts it as a mention.

Entity disambiguation — publishing does not touch it

From our own Turkish measurement: in 12.8% of answers that cite a brand's own site, the brand is never named (n=2,733). By engine it ranges from 3.4% to 25.6%.

Coverage

Five answer surfaces, reported separately

Engines do not agree. In our own benchmark one brand was named in 70% of Google AI Overview answers and 20% of ChatGPT's — same questions, same day. A single-engine reading would have made that brand look either excellent or hopeless.

We do not blend engine shares into one composite. No reliable platform-share data exists for Turkey, and a weighted score built on a number nobody has would be a fabricated metric.

same brand · same questions · same day

Google AI Overview%70
Perplexity%47
Gemini%40
Claude%33
ChatGPT%20

From our own benchmark, 27–30 questions per engine. 3 SERP calls failed and were excluded from the denominator.

Recorded for every question

Evidence, not a percentage

01The full answer text
Stored uncut — quotes, context and every later measurement depend on it.
02Whether you were named, and where
Opening sentence, bottom of a list, or after a 'however'.
03Who was recommended
The brand sitting first in that answer — usually a rival.
04Which sources were read
Every domain the answer rests on, including yours.
05How you were described
Positive, neutral, negative — and flagged separately if named as a warning.
06Whether it was you
If it was a namesake, that answer does not count toward your visibility.

Method

Unmeasured is not zero

The most expensive bug in this category is a measurement that failed and reached the report as 'not found'. A provider outage becomes 'this engine never mentions you', and a client budgets against it. Here every unmeasurable row is marked, excluded from the denominator, and retried.

  • Every failed call is stored with its reason — never silently skipped.
  • Every report states how many questions were actually measured.
  • A truncated answer is never counted as 'you were not named'.
  • An empty AI Overview is a real result: Google shows no box for that query.

Common questions

How many questions?

Between 30 and 250 depending on depth. Questions come from real buyer motivations rather than keywords, and the report shows every question asked and which persona it came from.

Do you scrape?

No. Every answer comes from the engine's own interface or official API and is stored in full. You can trace any sentence in the report back to its source.

How is the score calculated?

Five pillars: visibility, citation share, sentiment, content and technical. The weights and every formula are written inside the report — no black-box score.

How often is it measured?

Weekly or monthly. The question panel is frozen so trends stay comparable; if the panel changes, the report says so.

Find out which one it is

The free scan measures your site's AI readiness and gives you a first look at what the engines say about your brand.

Measure my brand free