How the score actually works
This is the real formula our scoring engine runs on every answer, for every prompt, across every engine.
Score each answer
Every time we ask an engine a prompt, we analyze the real answer text and score it 0-100:
Brand mentioned
CriticalYour brand appears anywhere in the answer
Recommended
The answer actively recommends your brand, not just mentions it
Position bonus
Rank 1 gets +20, decaying by 5 per rank, to 0 beyond rank 5
Sentiment
How positively or negatively your brand is framed
Cited domain
The answer includes a citation pointing at your own domain
* The five parts sum together and are clamped to 0-100 — a single answer can't score above 100 or below 0 no matter how the components add up.
Average per engine
Your score for ChatGPT, Gemini, Claude, Perplexity, Grok, or Google AI Overviews is the mean of every prompt's score run against that engine, plus how often you were named at all ("named in answers") as a separate stat.
One overall score
Your overall visibility score is the unweighted mean of all six engine scores — every engine counts equally today. Share of answer (how often you're named versus how often competitors are, across every answer) is tracked as a separate metric alongside it.
The verdict
Your overall score maps to one of three verdicts:
Invisible
(0-32)
Emerging
(33-65)
Strong
(66-100)
Week-over-week
Every scan compares your new overall score to your previous one — that delta is what shows up as your WoW trend everywhere in the product, computed the same way every time, never estimated.