Visibility#
The share of valid answers that mention your brand. A mention is the brand’s name in the answer text, not necessarily a recommendation. An AI Overview that Google did not show is not an answer.
Share of voice#
Your share of the mentions of every brand in the answers: the ones you monitor and the ones Lastfind discovers. Each brand counts once per answer. Names you rejected, names still in review and names that are a customer, supplier, partner, source or not a brand stay out.
Average position#
Where your brand first appears among the brands of an answer, averaged over the answers that mention it: 1 is first. Lower is better, and it only exists where the brand appears.
Citations#
A citation is a link the answer shows as its source. The citation rate is the share of answers that link to one of your domains (primary and additional, set in Brand and market).
Sentiment#
Each answer gives each brand it names a score from 0 to 100 on fixed criteria, starting at 50: promoter from 60 to 100, neutral from 40 to 59, detractor from 0 to 39. The sentiment index averages the questions, so a question with many answers does not outweigh the others. The bar shows the share of promoter, neutral and detractor answers.
Semantic Branding score#
Each AI gives each brand a note from 0 to 100 on each attribute; a note of 60 or more needs a concrete fact of that attribute. An attribute is the mean across the AIs, and a brand’s score is the mean of its attributes, each weighing the same. The score shown is steadied over the latest collections, because one AI’s note of a brand moves several points from one day to the next with no change in what the AIs say: each AI’s note is a weighted mean of its notes in the last 7 daily collections (3 weekly ones; monthly ones are not steadied), each collection weighing half as much every 2 days (every week). On the production projects this cut the day-to-day movement of a brand’s score two to four times. A real change still shows within a few days: about a third of it the next day, half after two days. The day’s own score is shown under the score, and the justifications keep each AI’s notes of the day. A brand with no note of its own in the period has no score, never one carried from before.
Topic weights#
Rates across topics are weighted averages. Set each topic’s weight in Settings > Monitoring > Topics: weight and models; a topic without a weight takes the smallest one.
A missing number is never shown as zero: when an AI did not answer or a report has no data for the period, the screen says so.