How we calculate GEO Score and where it falls short

Most AI visibility tools give you a number out of 100 and leave it there. We put our formula in the docs, so here it is in plain words.

GEO Score = Visibility Rate × 0.7 + Keyword Coverage × 0.3

Visibility Rate is intensity: citations divided by checks. Ten prompts across five assistants makes 50 checks. If you're cited in 12 of them, that's 24%.

Keyword Coverage is breadth: the share of your prompts where at least one assistant cited you. Those same 12 citations can come from 3 prompts or from 9. Coverage is 30% in the first case and 90% in the second.

Same visibility, very different picture. The first brand scores about 26, the second about 44.

Where it falls short

The score can sit still while things move underneath. If visibility rises while coverage drops, the number barely changes. That's why the dashboard always shows both parts next to it.

Citations are yes or no. Being the first recommendation in an answer and being a link at the bottom count the same.

That second one bothers me most. Should a top recommendation weigh more than a footnote link? And is 70/30 the right split, or would you weight breadth higher?