Benchmarks

AI Visibility → benchmarks ranks share of voice between you and your tracked competitors over the lookback window: one question — "who is winning AI answers in my category?" — answered as a picture instead of a research project.

The SOV viewCopy Anchor Link

• Brands (yours + tracked competitors) are ranked by mention share over the lookback window, with mention counts, average position and positive-sentiment split per brand.
• SOV is zero-sum: your gains are someone's losses. A rising share for a competitor is a falling share for you even when your absolute mention count is flat.
• The time series on this page is your own mention-rate trend; per-brand history and an engine breakdown of SOV are not available today — compare engines via the metrics page's per-engine rollups instead.

How to reactCopy Anchor Link

• A competitor pulling away — open the citation table: they are usually being cited more, not just named more. Compare the pages of theirs engines cite with your own for the same prompts; the structural difference is the playbook.
• New entrant appearing — add them to Competitors (with aliases) so SOV and every gap analysis starts accounting for them.
• Flat lines everywhere — your category's answers may be stable; shift attention to expanding the prompt set to new buying-journey stages instead of re-running the same questions.

Honest measurementCopy Anchor Link

Every run behind these numbers carries its data_source label, but the SOV aggregate itself blends live and sampled runs together (there is no live-only SOV view today — see metrics). When you report internally, caveat the blend: a share built mostly on sampled runs is a weaker claim than one built on live runs. See Data sources.