Ranking
The whole field on one page, ordered by whichever of the four figures you are asking about.
The Overview carries this table cut to its top six and ranked on visibility, which answers “how am I doing against the field” and stops there. This is the rest of it: every tracked brand, and the ability to ask a different question of the same rows.
Who leads on visibility is not who leads on sentiment, and a panel that is already a top-six cannot be asked which.
The columns
| Column | What it reads |
|---|---|
| # | Place in the current ordering, not a standing rank |
| Brand | Your brand and every rival on Brands |
| Visibility | Share of answers that name the brand at all |
| SOV | Share of all brand mentions that were this brand |
| Sentiment | How the answers talk about it when they do |
| Position | Where in the answer it tends to appear |
The four figures are all sortable, and the one you sort by is the question you are asking. A brand can be named in more answers than anybody and still be mentioned in passing at the end of each one — visibility and position disagreeing is a finding, not a contradiction.
A brand with no reading sorts last either way
A rival nobody has measured yet has no position, and an unmeasured brand is not a brand that scored zero. It goes to the bottom whichever direction the column runs, rather than topping a list of “best positions” with an absence.
Scope
The same bar as the Overview and Sources: date range, engine, country and topic.
Each filter is built from what has actually been sampled, so a project measured on one engine draws no engine filter. A filter with one thing to filter by is not a filter, and drawing it implies a choice that does not exist.
Narrowing the scope re-reads every figure in the table under it. Visibility on one engine is a different number from visibility across all of them, and it is the engine-by-engine spread — not the average — that usually explains a figure somebody is arguing with.
Where the numbers come from
This page reads the same query the Overview does, under the same scope. Two consequences, both deliberate: the two pages cannot disagree about a figure, and moving between them costs nothing because the cached result is the same.
Every figure is computed from the answers on Chats. A number here that looks wrong is traceable to the answers behind it, which is the point of keeping them.