The GeoNexo scoring model — what our visibility % means

TL;DR — Your visibility score is mentions ÷ responses × 100. One formula, applied consistently across every engine and every prompt, so the numbers are directly comparable. No rank-tracker "difficulty" scores, no black-box weightings.

The formula

For any slice you look at — a single prompt, a single engine, a competitor, your whole brand — the score is calculated the same way:

visibility = (mentions ÷ responses) × 100

• mentions = the number of AI responses (across engines × prompts × days in the window) that named your brand
• responses = the total number of AI responses in that window

That's it. Everything else in the dashboard rolls up from this one number.

Why we use one formula everywhere

A common frustration with rank trackers is that "your score" means different things on different tabs — one page uses average position, another uses share-of-voice, another uses estimated traffic. You end up not trusting any of them.

We chose the opposite direction: one formula, applied everywhere. Whether you're looking at a single prompt on Gemini this morning, or your brand across all engines over the last 90 days, the arithmetic is identical.

The trade-off is honesty over sophistication. We could weight bigger engines more, we could adjust for citation quality, we could layer semantic similarity — but every added variable is another thing you'd have to trust us on. Simple wins.

Reading a score

Rough bands for a typical B2B brand in a competitive category:

Score	Meaning
0–5%	You're invisible in AI answers for this slice
5–15%	Occasional mentions, mostly circumstantial
15–30%	You're on the shortlist
30–50%	You're a default answer
50%+	You dominate this slice

These bands are directional, not gospel. Different categories behave differently — a niche vertical with three real players will run higher than a saturated one with dozens.

What we compute on top of the score

The one formula gives you the raw number. Everything else in GeoNexo is a lens on it:

• Per-engine breakdown — same formula, filtered by engine
• Per-prompt breakdown — same formula, filtered by prompt
• Competitor visibility — same formula, applied to competitor mentions instead of yours
• Trend over time — the daily score plotted across the window you pick
• Gap prompts — prompts where your score is zero but a competitor's isn't

What we do NOT expose (and why)

We keep the internal prompt-selection logic, competitor de-duplication rules, and the engine-specific mention detection heuristics private. Not because they're mysterious — because they're the moat. Publishing them would let a competitor game the score without moving the underlying reality. Every score on your dashboard reflects a real AI answer we captured that day.