Share of Voice
Share of voice is the comparative view: across your tracked prompts, how much of the available attention did each brand capture?
What goes into it#
Mention frequency. How many of the tracked answers named the brand at all.
Position. Where in the answer. First is a recommendation; last in a list of eight is barely presence.
Sentiment. How the brand is characterised. "The most complete option" and "a cheaper alternative with fewer integrations" are both mentions and they sell very differently.
Reading the comparison#
A few shapes recur:
You are absent, they are present. A pure source problem. Go to the citation list for those prompts and find what carried them.
Both present, they rank higher. Positioning. The model has a reason for preferring them, and it is usually visible in the answer text.
Both present, they read better. A sentiment problem. Something in the cited sources describes them more convincingly than you.
You lead one group, trail another. The most common and most useful result. It localises the work to a specific buyer stage.
Track it over time#
A single reading is a snapshot of a probabilistic system. The trend is what tells you whether last quarter's content push actually moved the needle, and whether a rival's push is moving theirs.