Read your three metrics
What Siggmund measures
Every week, the Siggmund panel asks five AI engines (ChatGPT, Gemini, Google AIO, Google AIM and Perplexity) the questions your customers really ask, then reads their answers. Three metrics come out of it. On the dashboard, the engines are combined by their median.
They appear in the application as Recommendability Score (the "Score" in the tabs and reports), Citation rate and Leading rate.
Citation rate
The share of answers in which AI assistants cite your brand.
It is the simplest measure: either the AI cites you, or it does not. A Citation rate of 30% means you appear in three answers out of ten.
Leading rate
Among the answers that cite you, the share made with a strong purchase intent.
Being cited in "what is an e-bike" and being cited in "where to buy an e-bike in Lyon" are not worth the same. The Leading rate tells them apart: it rises when you are cited at the moment the person is ready to act.
Recommendability Score
The Recommendability Score combines the two: it is the Citation rate multiplied by the Leading rate.
It answers a single question: does your visibility bring you customers? A brand cited often, but never at the right moment, will have a low Score, and that is intended.
How to read them
- Look at the gap, not the absolute value. Each sector has its own orders of magnitude. On the Performance page, gaps are shown in points against the median.
- Follow the trend over several weeks. A single week says little: AI answers vary from one time to the next.
- Go down to the quadrant level. A stable average can hide a sharp rise and a sharp fall that cancel each other out.
What these numbers do not tell you
- They do not measure your traffic, but your presence in AI answers.
- They do not count the engines Siggmund does not query.
- They rest on a sample of questions: there is a margin. In a quadrant's detail, the Score curve comes with its 90% credible interval. When that interval is too wide, Siggmund shows the range only, without a central value.
Last updated: · First published:
