Paid SOV Tracker plans compare your brand with a configured competitor universe across five engines. Canonical SOV stays a pure competitor-relative mention share; recommendation, citation, sentiment and prominence evidence remain separate so you can see what actually changed.
The comparison universe and measurement scope stay explicit so one score cannot hide incompatible evidence.
01
On paid tracking plans, add the competitors that belong in the comparison set for your brand. Canonical SOV is only meaningful when that comparison universe exists; the Free Scan does not manufacture competitor-relative SOV.
02
During compatible synchronized measurement windows, SOV Tracker evaluates the returned answers for the target brand and configured competitors on the same tracked question scope. It counts valid mentions inside the evidence returned by each engine rather than treating separate brand-specific questions as equivalent evidence.
03
Compare raw Share of Voice, valid mention counts, citation evidence, Sentiment and Prominence as separate readings. A prominence or sentiment signal never gets blended into the SOV percentage.
04
Track movement only across compatible dated measurement windows. AI answers can change with engine, model version, locale, question and time, so a before/after comparison should preserve the scope instead of implying causality from one content or PR change.
Use the evidence to describe the measured difference, not to infer an unobserved cause.
See the target brand's valid mention share inside the configured comparison universe. Canonical SOV is the target's valid mentions divided by valid mentions for the target plus configured competitors on the compatible scope.
Inspect which engines returned the brand or competitors, the answer evidence available to the scan and observed citation URLs. Engine identity stays explicit instead of being hidden behind one blended score.
Prominence shows where the brand's first valid mention appears inside the returned answer. It is not a recommendation rank, and it remains separate from SOV.
Provider errors, timeouts and other technical failures remain not measured. They are never rewritten as a competitor win, a brand loss or an artificial zero. When target and configured competitors have zero valid mentions, SOV is unavailable (0/0), with a distinct reason from provider failure. Measured 0% SOV requires valid competitor mentions and no valid target mentions.
SOV is the share metric. The other readings explain different parts of the returned evidence and are not ingredients of one blended composite metric.
Competitor-relative percentage from valid mentions on the compatible configured scope.
Raw valid occurrences for the target brand and configured competitors.
Observed source URLs in the evidence available to the scan. A cited source is not a recommendation.
A bounded positive, neutral or negative classification of the analyzed response.
Where the first valid brand mention appears inside the returned answer; not a market or recommendation rank.
Free is a five-question, non-branded ChatGPT Discovery scan and does not produce competitor-relative SOV. Paid plans add configured competitors, tracked questions and synchronized five-engine measurement.