Share of voice

Share of voice is your proportion of the brand names models hand back across your tracked prompts. It is relative by definition, so the number says as much about your competitor set as about you.

How it is counted

For every run, the brands named in the answer are extracted and mapped through Name matching. Your share is your appearances over the appearances of the whole mapped set, across the prompts and the range you have selected.

Two consequences worth internalising:

• Adding a competitor lowers your share without anything changing in the market.
• A brand nobody has mapped is invisible to the metric, so an unmapped rival makes your share look better than it is.

Rank, not just share

Alongside the share, the owned brand carries a rank: how many brands are named before yours. Two brands can hold the same share while one is consistently named first and the other consistently named last, and only one of those positions wins a deal.

Watch rank when the share is flat. It often moves first.

What to compare it against

Itself, over time, with the competitor set held still. If you change the set, note the date, because the series before and after it are two different measurements.

Reporting it

Three things make a share-of-voice slide hold up.

• The prompt set, stated. "Across 24 buyer prompts" beats a bare percentage.
• The competitor set, named. The reader needs the denominator.
• The range and the cadence, so they know how many probes are behind it.

When it is the wrong metric

If you are the only brand models know in your niche, share of voice reads near 100 percent and tells you nothing. The Visibility Score is the better number then: the question is not whether you beat rivals, but whether the model recommends anyone at all for that prompt.

The reverse case: in a market with forty credible vendors, a 5 percent share can be a strong position. Read it next to rank before deciding it is bad.