Answer Radar
Adding Competitors to Answer Radar
Competitors are added on the project form, one name at a time, and they do two jobs: they define whose visibility is measured beside yours, and they let a finding say which competitor an engine recommended instead of you. This article explains how tracked and untracked competitors differ, and how vendors and products are reconciled so a broad catalogue cannot inflate anyone's score.
Quick Answer
Add competitors as brand names on the project form. Tracked competitors get share of voice, mean position and sentiment beside yours. Competitors you did not add are still detected from answers and raised as competitor won cards.
- Share of voice counts a vendor once per answer, however many of its products the answer lists.
- Only products count; publications and features named in an answer are excluded.
- Vendor identity is reconciled across the whole run before anything is measured.
On this page
Tracked versus untracked competitors
A tracked competitor is one you named on the form. It is a product by definition, whatever the mention judge says, so it can never drop out of the metric that measures it. Its share of voice, mean position and sentiment mix are reported beside your brand's in every segment.
An untracked competitor is a brand an engine recommended that you did not name. Answer Radar still detects it and raises a competitor won card keyed on what was observed, a domain or an entity name. If you later add it as tracked, its evidence starts carrying a row id and the derived key changes, so the finding is raised again as a new card while the old one keeps its claim until dismissed. That fails in the direction of showing one extra card, never of hiding a finding.
One vendor, counted once
Every judged mention carries a kind, product, publication, feature or other, and a vendor key. Only products are admissible to share of voice: a review site the answer cites and a feature the answer discusses are not candidates you compete against. A deterministic string rule rolls product lines up to their vendor first, so a brand named beside two of its product lines is one vendor, and the judge's answer is used only for parentage no string can reach.
Reconciliation runs across the whole run before segmentation, because 'this product belongs to that company' is a fact about the world and applies whichever engine revealed it. It may only merge, never split, so a vendor left split is understated rather than anyone being credited with visibility they did not earn.
How competitors show up in results
In metrics: share of voice per vendor summing to one across vendors, mean first-appearance position per vendor, and sentiment mix per vendor, each per segment. In gaps: the observations name who was recommended and the citations show which pages put them there. In the action queue: competitor won cards carry impact counts such as competitor citations and competitor mentions, with the citation ids behind them.
A vendor key is an attribution, never a presence claim. Nothing in vendor grouping decides whether your own brand was present in an answer; that is a separate whole-word match against the answer text.
- Name the brand, not the legal entity or the product line.
- Five or fewer keeps share of voice readable.
- You cannot inflate or deflate a competitor by adding product lines; the sample is the unit.
FAQ
Why does a competitor I never added appear in my gaps?
Because an engine recommended it. Detection works from what the answer said, not from your list. Add it as tracked if you want its share of voice measured beside yours.
Why is a review site not counted as a competitor?
Entity kind. Publications and features are excluded from share of voice on purpose. A site the engine cited is a source, and sources are reported under citation share instead.
Can I track a competitor by domain instead of name?
The form takes a brand name. Domains appear in citation share, which flags whether a cited domain is yours or a tracked competitor's, so a competitor's own pages are attributed once its name is known.
Related help pages
Answer Radar
Creating an Answer Radar Project
What each field on the new project form does, why only the domain is required, how idempotent creation protects a retry, and how archiving reverses safely.
Answer Radar
Reading the Visibility Metrics
What presence rate, recommendation rate, share of voice, sentiment mix and citation share mean, which samples count, and why a segment can report no number.
Answer Radar
The Action Queue
What the Actions tab holds: content gap and competitor won cards derived from measured runs, their evidence and impact counts, the five statuses, and why no.
Answer Radar
Reading the Mean Position Metric
What mean ordinal position measures, why it comes from the answer text rather than the model's ranking, why absence is not zero, and why it is never subtracted.