Skip to content
New: already collect AI answers? Bring them, and we score them
Visibility Monitoring
Home / Approach / Reading the results
Our approach

How to read the results

A score is only useful if you know when it really moved.

Three rules cover most of it: ignore small swings, give fixes time, and open the answers.

Illustrative data. Answers naming you rose from 46 percent in run 8 to 52 percent in run 9, and the move holds with any one AI tool left out, so it is called real. By tool: ChatGPT 33 to 35 percent, within noise; Gemini 51 to 58, a real move; Perplexity 47 to 49 and Google AI Mode 29 to 27, within noise.
Example. A move is called real only when it holds.
Rule 1

Small swings can be noise

Example, illustrative data: a line chart of twelve weekly runs of the Live Visibility Score against a dashed baseline, with a logged change marked at week 7.
Illustrative. A logged change is marked on the trend, so you can see what followed it.
  • How far a reading moves on its own depends on the metric, the number of answers and the AI tool, so no single number of points applies to every reading.
  • In testing, three asks on one day disagreed in about 1 in 7 question and tool pairs. Why we ask three times
  • The platform tests a move against that variation before calling it real.
  • A move must also hold without any single AI tool, or it is labeled one tool only.
Rule 2

A change is checked, not assumed

Day 0
You publish

Log the change on the platform.

Next runs
AI tools re-read

The time this takes differs by platform and source. No fixed delay has been established.

Before and after
Same questions compared

Measured against your baseline. A move after a change is an association, not proof that the change caused it.

Rule 3

Open the answers behind the number

Illustrative data. Brand A is visible 100 percent of the time on ChatGPT and absent elsewhere: fragile. Brand B is visible 20 percent of the time on all six platforms evenly: durable.
Illustrative. Two businesses with similar totals can stand on very different ground.
What a score cannot prove

That AI caused a booking or a sale. Connect your analytics for traffic AI sent you.

What it can show

What AI said, to whom, on which tool, and how that changed after your work.

See the approach on your own market.

Start full monitoring, where we ask the AI tools for you, or upload answers you already collect.

Start monitoring Import your own answers