Identify Competitors Cited Instead of Your Brand
Track which domains win the citations your brand does not receive, prompt by prompt, using Content Hub MCP in Claude Desktop.
Your own citation rate shows only part of your position. To prioritize work, identify which competitors win the citations you miss, on which prompts, and whether the gap is widening or narrowing. With Claude Desktop connected to Content Hub MCP, you can retrieve this analysis in a single conversation instead of comparing domains manually.
Before you start
- Create a GEO cluster and add prompts to it.
- Complete enough runs to provide evaluation data for comparison.
Fetch the cluster overview
Request the detailed cluster data for the Developer Marketing Agency cluster in the Infrasity organization, and name Content Hub MCP as the data source.
Claude resolves the company ID and then the cluster before it returns a summary. The summary reports the following:
- Infrasity is cited across most prompts and holds the #1 rank on the majority of them.
- The strongest prompt, "best developer marketing agency for AI startups", has an 88% citation rate and 15% share of voice.
- The most frequent competitors across the cluster are hackmamba.io, draft.dev, catchyagency.com, devmarklab.com, dev.to, and business.daily.dev.
The overview also lists the weak spots. Six prompts have a 0% citation rate, and each is named individually. Examples include "How to measure ROI in developer marketing" and "Developer marketing agency pricing and packages". These prompts indicate a content gap rather than noise. A rank-movement summary follows and identifies the pricing and ROI prompts as the largest opportunity.
Review the win/loss breakdown by prompt
Request a detailed breakdown of where Infrasity wins and loses citations, by model and by prompt, for the same cluster.
The underlying API aggregates results across all four tracked models: Gemini, Claude, Perplexity, and ChatGPT. It does not expose a per-model split, so every figure is blended across engines.
The first table lists the top wins. These are prompts with a citation rate of 80% or higher and a rank of #1 or #2. Each row includes the runner-up domain and its share of voice. For example:
- "best developer marketing agency for AI startups" has an 88% citation rate. Infrasity holds 15% share of voice, and the runner-up holds 11%.
- Several "devtools" and "API startup" prompts have an 84% citation rate and 12% share of voice each.
The second table lists prompts that improved in rank. Each row shows the rank change, the current citation rate, and the competitors that still lead. One prompt moved from rank 6 to rank 1 but still trails two named competitors on citation share. A rank gain and a share-of-voice gain are therefore separate outcomes.
The third table lists prompts that declined. Each row shows the rank change, the citation rate, and the domains overtaking the prompt. This table exposes specific threats that a cluster-wide average hides. The zero-citation table follows. It pairs each gap with the domains that dominate it and the share they hold.
The final section summarizes the cluster:
- A competitor frequency table ranks each recurring domain by the number of the 40 prompts in which it appears. Each row describes the competitor's focus. One domain dominates definitional and "how to choose" prompts. Another is strong on startup and Series A prompts.
- A trend shows the cluster citation rate over the full 30-day window.
- Infrasity wins on service-category, API-specific, and startup-segment prompts. It has no citations on commercial prompts such as pricing, cost, ROI, and packages. It is losing ground on definitional and comparison prompts to the same two competitors.
Request takeaways
Ask Claude to convert the win/loss data into takeaways. Claude returns a numbered list. The first two takeaways are as follows:
- Intent split. Infrasity dominates intent-specific searches, often ranking #1 with double-digit share of voice on "best agency for X type of startup" prompts. Every prompt about pricing, cost, packages, or ROI has a 0% citation rate. This is a content-existence problem rather than a ranking problem. No page on infrasity.com appears to cover pricing or ROI, so models have nothing to cite.
- Priority competitors. One domain is the default explainer and directory source for definitional and "how to choose" prompts. A second domain recently overtook Infrasity on a prompt it held one cycle earlier. Both warrant priority over smaller sites that gain only a few points of share of voice.
The remaining takeaways are as follows:
- Rank drop. The specific rank drop is the most actionable signal. Check whether Infrasity's comparison content is stale or whether a competitor published fresher content.
- Improving prompts. Identify what changed in the content that improved, and apply it to the declining and zero-citation prompts.
- Citation-rate trend. Verify the late drop in the overall trend. It may be a regression or a low-sample day, and the distinction determines the urgency of the response.
Claude also suggests a priority order:
- Write the pricing and ROI content, because it closes the most gaps at once.
- Refresh the page that lost its top rank.
- Audit and replicate what works on the improving prompts.
- Monitor the citation-rate dip to confirm that it is real.
A single Claude Desktop conversation produces a cluster overview, a win/loss breakdown, and a prioritized action list. Each competitor threat traces to a specific prompt and rank.
To review the competitors beating your brand, open app.infrasity.com and retrieve the win/loss breakdown for your own cluster.
