What Share of Voice means
In classic SEO it's usually computed from ranking positions weighted by search volume. In AI search it's simpler: across a fixed prompt set, what proportion of answers name you, and what proportion name each competitor.
It's more useful than an absolute citation count because it accounts for how contested the category is.
Why it matters for AI search
Absolute citation numbers can rise while your position worsens, if competitors are gaining faster. Share of voice catches that; a raw count doesn't.
It's also the metric that makes sense to executives, because it frames visibility competitively rather than as an isolated number.
See it in action
Your citation count rose. Your share fell.
Everyone gained. Two competitors gained faster.
Your citations went up, which looks like success in isolation. Share of voice shows you lost ground relative to the category, which is the number that actually predicts pipeline.
How to get it right
Measuring share of voice
- Define a fixed competitor set — three to five named rivals, not a vague market
- Run the same prompt set for you and them, in the same cycle
- Report proportion, not raw counts, so category growth doesn't flatter you
- Track the 'nobody named' share too; it shows how much territory is still unclaimed
- Re-baseline the competitor set annually rather than mid-year
Common questions
How is this different from citation share?
Citation share is your proportion of the prompt set. Share of voice adds competitive context, showing your slice relative to everyone else's.
How many competitors should we track?
Three to five. More becomes noise; fewer misses the rival who's actually gaining.
What if nobody is being cited in our category?
That's genuinely useful information — an unclaimed field is the cheapest time to establish position. It also might mean your category isn't asked about conversationally yet.
Related concepts
These come up alongside Share of Voice constantly.