Platform · Share of Voice
See who wins the conversation, month by month
Share of voice per model against a declared competitive set, with mentions and citations kept apart.
The brands we measure every day
The metrics
From the aggregate share to the single answer
Mentions and citations kept apart, a monthly reading, a matrix by cluster and drill-down all the way to the text of the answer.
-
Share of voice by model
The share is not a single number: it changes by model, and disagreement between models is operational information.
Share of voice last 30 days -
The figure is read on the month
We collect every day, but the reference reading is over windows of at least thirty days: below that, you measure noise.
Share of voice last 12 months72% da 41%- Oct
- Jan
- Apr
- Sep
-
The brand × cluster matrix
Aggregate share of voice tells you whether you are winning. The matrix tells you where, and where nobody holds the conversation.
Brand × cluster mock-upComparison Choice Price Usage Your brand 78 54 22 6 Competitor A 64 71 40 18 Competitor B 31 28 66 24 No one 0 0 14 62 -
Mention and citation
A mention says you exist in the answer, a citation says the model trusts your source.
Answers that include you mock-up1.284 answers- Mention only 64%
- Mention + citation 36%
-
Alerts on the band of variation
The threshold is not fixed: it is estimated on your own prompt set, so an alert flags a real movement.
Alerts this period mock-up- Comparison cluster · ChatGPT -6,2 pt Out of band
- Price cluster · Perplexity +3,4 pt Relevant
- Usage cluster · Gemini +0,8 pt Noise
Being named is not being recommended: a mention says you exist, a citation says the model trusts you.
Frequently asked questions
Taken verbatim from the monitored prompts. FAQPage schema.
How is share of voice calculated?
As the share of answers that mention your brand out of the answers relevant to the prompt set, by model and by cluster, against a declared competitive set.
Why not a single cross-model number?
Because the models disagree, and an average would delete exactly the information that tells you where to act.
Is the competitive set fixed?
No: it is declared at setup and can be revised. Every change stays on the record so the historical series remains readable.
How long before the figures are meaningful?
At least thirty days. Below that window daily variation dominates the signal.
Try it on your own data
Start the 7-day free trial yourself (card required, no call needed): see your real data before choosing a plan.
