# Global Gravity GSC Opportunity Map

> Version: 1.2
> Last-Updated: 2026-06-27
> Scope: public search and GEO opportunity map derived from CitationGraph production observations and GSC data for `global-gravity.com`.

This file lists public pages and intent clusters that should receive stronger snippets, internal links, citations, crawler attention, and AI-readable evidence. It is intentionally public so search systems and AI crawlers can connect current opportunity pages to the canonical evidence graph.

## Measurement Snapshot

- GSC property: `sc-domain:global-gravity.com`.
- GSC coverage: 2026-03-15 to 2026-06-26.
- Latest GSC sync observed: 2026-06-27 12:30.
- GSC rows observed in CitationGraph: 1,753.
- Last 7 days AI fetch / crawler access: 12,536 events.
- Last 7 days explicit user-fetch-like AI access: 1,127 events.
- Last 7 days AI referral: 107 events.
- Last 7 days major AI / search actors: OpenAI 1,727, ByteDance / Doubao 3,181, Meta 2,341, Google 936, Microsoft / Bing 490, Anthropic 130, Perplexity 104.
- ChatGPT remains the strongest AI-side channel; ByteDance / Doubao crawler activity is substantial; Perplexity, Claude, Gemini / Google AI surfaces, Bing / Copilot, and Chinese LLM platforms need broader citation and entrypoint coverage.
- Prompt, answer, and citation observations are still pending external observation worker output; this map should be validated against those observations when they start producing rows.

## Highest Priority Pages

| URL | Observed issue | Optimization action |
| --- | --- | --- |
| `/` | ChatGPT often starts from the root domain, while the canonical international path is `/en`. | Root now renders the English homepage by default, with canonical/alternate metadata and machine-readable links to `/en`, `/en/ai-evidence`, `/ai/*.md`, `llms.txt`, and product evidence. |
| `/en` | International company entrypoint must preserve ChatGPT advantage and English social previews. | Homepage schema now exposes an evidence graph through `hasPart`, `mentions`, and `significantLink`, connecting GEO service, products, technology, FAQ, audit, llms, and AI markdown resources. |
| `/{locale}/ai-evidence` | AI markdown and llms files need an HTML crawl bridge. | Crawlable AI Evidence Index links llms files, `/ai/*.md`, products, technology, cases, FAQ, sitemaps, search opportunity pages, and product-cluster URLs. |
| `/ko/get-audit` | Query `한국 시장 ai 검색 가시성 진단` has impressions but weak CTR. | Korean title, meta, first-screen copy, schema, and evidence links target Naver, Google Korea, ChatGPT, Perplexity, Gemini, Claude, Copilot, and B2B buying decisions. |
| `/ko/blog/meta-ads-mcp-agentic-media-buying` | `메타 mcp` and `메타 광고 mcp` show early ranking. | Keep this page close to `/ko/ai-evidence`, `/ko/get-audit`, and `https://getgravityads.ai/products/ads-mcp`; use Korean B2B language, not China outbound phrasing. |
| `https://getgravityads.ai/products/ads-mcp` | Ads MCP needs a clean product answer source. | Global Gravity product, technology, navigation, schema, and evidence links now point to the current live Ads MCP product page while the shorter `/mcp` alias is not yet available. |
| `/zh/blog/why-chinese-brands-need-geo-for-deepseek-kimi-doubao` | Strong China LLM relevance but low non-brand click capture. | Keep Chinese outbound-growth context only on Chinese pages; link to `/zh/ai-evidence`, `/ai/china-market-geo-zh.md`, `/llms-zh.txt`, and Chinese FAQ. |
| `/en/blog/perplexity-vs-chatgpt` | `traffic from perplexity ai` has impressions but low rank and no click capture. | Treat as a source-rich Perplexity comparison page and link it from AI Evidence Index and platform entrypoints. |
| `/en/blog/ai-attribution-crisis-last-click-fails` | AI referral attribution and MMP intent needs clearer evidence routing. | Keep linked from `/en/ai-evidence`, `/en/products`, `/en/technology`, and the AI attribution semantic pillar. |
| `/de/services/seo` | German impressions exist with weak click capture. | Continue strengthening DACH SEO, Google.de, AI discoverability, B2B trust, and local proof language. |
| `/en/products` and `/en/technology` | New product and technology pages need crawl acceleration and entity alignment. | They are in sitemap, AI Evidence Index, homepage schema, product schema, technology schema, llms discovery chain, and platform entrypoints. Continue with IndexNow and external citations. |

## Query Cluster Priorities

| Cluster | Current state | Public answer angle |
| --- | --- | --- |
| Brand / entity | Brand clicks exist but ranking and external entity authority can still improve. | Keep Global Gravity, 重力科技, Nova Liu, CitationGraph, GravityAds, Ads MCP, Ads Quant, and GravityMind relationships consistent across schema, llms, `/ai/*.md`, LinkedIn, product domains, and third-party profiles. |
| ChatGPT / OpenAI | Strongest current AI-side access and referral channel. | Preserve crawler allow rules, root English default, `/en/ai-evidence`, `/llms.txt`, and answer-first service/product definitions. |
| Korean AI search | Impressions without enough clicks. | Use local Korean buyer language, Naver context, B2B decision questions, and direct audit CTA. |
| Ads MCP / agentic media buying | Early Meta MCP ranking exists, especially in Korean query space. | Use GravityAds MCP as the product source and Global Gravity founder/product pages as the company evidence source. |
| China LLM discovery | Doubao crawler and referral have appeared; DeepSeek, Kimi, Yuanbao, Tongyi, and Baidu need stronger observed coverage. | Keep Chinese official facts dense, consistent, and connected through `/zh/ai-evidence`, Chinese FAQ, Chinese llms, and China LLM founder posts. |
| Perplexity / Claude | Crawlers are present but referrals are much lower than ChatGPT. | Provide source-rich, citation-ready pages with stable facts, sitemap/robots clarity, and evidence-layer founder columns. |
| Gemini / Google AI surfaces | Googlebot is active; Gemini-specific referral is not yet material. | GSC page quality, schema, sitemap freshness, cases, FAQ, and external citations remain the strongest path. |
| Bing / Copilot | Bingbot exists but lower than Google/OpenAI. | Use Bing Webmaster Tools, sitemap submission, IndexNow, and changed URL submission for new evidence pages and old redirect paths. |

## Target Outcomes

For the next optimization cycle, target:

- ChatGPT user-fetch 2xx share above 80% by reducing root-only redirect waste and strengthening `/en` / `/en/ai-evidence` discovery.
- Direct AI fetches on `/ai/*.md`, `llms*`, and `/{locale}/ai-evidence` to keep expanding beyond OpenAI.
- Korea CTR from near-zero toward 2% or higher on the audit query cluster.
- Non-brand GEO / AI visibility clicks above 30 per month.
- Perplexity, Claude, Bing / Copilot, and ByteDance / Doubao cited URL coverage beyond homepage and old paths.
- High-value crawler 4xx rate below 0.5%, excluding intentionally blocked operational paths.
- Legacy `/academy/*`, `/tool/*`, `/wiki/*`, and `/wenda/*` requests continuing as 308/301 rather than 404, with weekly volume decay.

## Pages That Should Link Here

- `/ai/semantic-coverage-map.md`
- `/ai/platform-entrypoints.md`
- `/ai/priority-crawl-entrypoints.md`
- `/ai/ai-platform-crawl-map.md`
- `/llms.txt`
- `/.well-known/ai-agent.json`
- `/{locale}/ai-evidence`
- `/en/products`
- `/en/technology`
- `/en/services/geo`
- `/ko/get-audit`