Head-to-Head Win Rate Matrix vs Competitors

Why a matrix beats a leaderboard

A leaderboard says you are #3 in your category. It does not say who is #1, who is #2, what they win on, and where you flip from beating them to losing to them. That is the work. The Lumos head-to-head matrix is built to surface those flips — the specific buyer prompts where competitive substitution happens.

In our own data, 78% of competitive prompts produce a clear win/loss verdict; the rest tie. Across 2,689 verdicts we have scored on customer brands, the median brand wins fewer than half of its category prompts against its biggest competitor. The bad news: most teams cannot tell which half. The good news: once you can see the matrix, the content roadmap writes itself.

What you see inside Lumos after signup

The preview above is Lumos's own head-to-head matrix, scored against our four most-mentioned GEO competitors (Profound, Peec AI, Otterly, Brandwatch) on six category prompts our buyers actually ask. After signup, the same surface runs with your competitor list and your prompts:

• Rows are buyer-intent prompts. Columns are competitors. Each cell is a verdict from the latest scan.
• Green = win (you are cited first, with positive framing). Red = loss. Gray = tie. White = not yet scored.
• Click any cell to open the raw AI answer that produced the verdict, the engine that returned it, and the cited sources.
• Filter by engine, time window, or buyer segment in the toolbar.

The matrix updates as the daily scan runs. A flip from win to loss triggers an alert if you have the notification add-on.

How Lumos scores a win

Per prompt, per engine, we parse the answer and look at three things in this order: which brand is cited first, in what context (recommendation vs comparison vs anti-recommendation), with what sentiment. A first citation with positive framing is a win. Last citation, comparison-only framing, or negative qualifier is a loss. Same rank, same framing is a tie. Aggregating across engines gives the per-prompt-per-competitor verdict you see in the matrix.

Common questions

What if my competitor is mentioned but I am not? That is a loss. Lumos does not soften the verdict with euphemisms.

Can I exclude prompts I do not care about? Yes. The matrix only shows the prompts you tracked. You define the slate; Lumos does the scoring.

Does the matrix include sentiment? Each cell carries a sentiment qualifier you can toggle on. For deeper language analysis, use Brand Sentiment Snapshot.

How often do verdicts flip? In our customer data, the median brand sees roughly 10–15% of cells flip week-over-week — enough that monthly check-ins miss the action.

Next steps

• Build your own matrix: sign up, pick four competitors, define six prompts.
• See your full visibility picture: Multi-LLM Brand Visibility Comparison.
• Auto-profile your competitors: Competitor Profile Builder.
• Read the strategy: GEO vs SEO and Why GEO is a competitive advantage.