Custyle

Why: the page HTML contains almost no text — content only appears after the browser runs JavaScript, and AI crawlers do not run JavaScript, so they see an empty page.
Where/what, pick your stack: Next.js → make the page a Server Component (remove "use client" from page-level components) or add static generation; Nuxt → ensure ssr: true in nuxt.config.ts (it is the default — check nobody set it to false), or prerender marketing routes with routeRules { "/": { prerender: true } }; plain Vite/CRA SPA → prerender marketing pages to static HTML (vite-ssg, react-snap) or move them to a static-site generator; Webflow/WordPress/Shopify already render server-side and rarely trigger this.
Check: run curl -s https://custyle.ai/page | grep "a sentence from your visible copy" — the sentence must appear in the raw HTML. Then: npx fastergeo verify.

Why: sentences with concrete numbers are the single most-cited content type in AI answers.
Where: anywhere in the main body of the affected pages (top half is better).
What: 2–3 sentences each containing a specific number plus its source/date. Example: "In an internal benchmark (June 2026, n=120), setup time dropped 43%, from 21 to 12 minutes." Vague claims ("many users", "significantly faster") do not count — the number must be a digit.
Check: npx fastergeo verify — the statistics block is detected automatically.

Why: when users ask category questions without naming you, engines rarely bring you up. This is a rollup metric — it moves through the tickets it spawned, not by itself.
What: 1. Complete the earned-media ticket below (third-party presence drives ~84% of citations). 2. Complete the on-page block tickets so your pages are quotable when engines do look. 3. Publish 1–2 pages answering the exact category questions in your questions.json (the same file the sampling runs on).
Check: nothing manual — next-period sampling (npx fastergeo cycle) remeasures the rate and flips this automatically. Move takes weeks, not days; judge by trend.

Why: when users ask category questions without naming you, engines rarely bring you up. This is a rollup metric — it moves through the tickets it spawned, not by itself.
What: 1. Complete the earned-media ticket below (third-party presence drives ~84% of citations). 2. Complete the on-page block tickets so your pages are quotable when engines do look. 3. Publish 1–2 pages answering the exact category questions in your questions.json (the same file the sampling runs on).
Check: nothing manual — next-period sampling (npx fastergeo cycle) remeasures the rate and flips this automatically. Move takes weeks, not days; judge by trend.

Why: AI answers "what is X" by lifting a definition sentence. Pages without one are skipped.
Where: the first paragraph right under the page H1 (must be in the top of the page, not the footer).
What: one self-contained sentence in the pattern "X is a Y that Z." Example: "FasterGEO is an open-source toolkit that measures whether AI assistants mention and cite your brand." Name the subject explicitly — do not start with "It is…".
Check: re-run your audit (npx fastergeo verify) — the definition block is detected automatically.

Why: 44.2% of AI citations extract from the top 30% of a page — your answer blocks sit below that line.
Where: the affected pages; this is a MOVE, not a rewrite.
What: cut the definition sentence and the key numbers currently buried mid-page and paste them directly under the H1 as the opening paragraph ("answer-first" order: conclusion → evidence → detail). Long intros, hero carousels and menu boilerplate push answers down — trim them.
Check: npx fastergeo verify — block position is remeasured.

Where: a "FAQ" section at the end of the affected pages.
What: 3+ real question-headings (H2/H3 ending in "?") each followed by a 2–4 sentence answer. Write questions the way buyers ask them ("Is X free?", "Does X work with Y?"). Optionally add FAQPage JSON-LD with the same Q/A pairs.
Check: npx fastergeo verify.

Where: main body of the affected pages.
What: an HTML <table> (or Markdown table) comparing your product with 1–2 named alternatives across 3+ concrete criteria (price, deploy model, license…). One honest row where a competitor wins increases credibility and citation odds. A styled <div> grid does NOT count — crawlers extract <table> markup.
Check: npx fastergeo verify.
Competitors AI actually names in sampling: Etsy, Printful, Zazzle — compare against these first.

Where: main body of the affected how-to/setup pages.
What: a numbered list (<ol> or Markdown "1. 2. 3.") of 3+ imperative steps, each starting with a verb: "1. Install the CLI: npm i -g fastergeo. 2. Run npx fastergeo audit --root https://custyle.ai. 3. Open report.html." Prose like "first you should…" does not count — it must be list markup.
Check: npx fastergeo verify.

Why: pages under ~600 word-equivalents rarely satisfy an AI answer on their own.
Where: the affected pages (listed in the audit report).
What: extend each to 600+ words by adding the missing sections rather than padding: a definition paragraph, a numbers/evidence paragraph, a steps list, and a 3-question FAQ. That structure alone typically adds 300+ useful words.
Check: npx fastergeo verify — word-equivalents are recounted on re-crawl.

Where: inside <head> of each affected page template (one <script> tag; in most frameworks add it to the page layout/SEO component).
What: paste and adapt:
  <script type="application/ld+json">
  {"@context":"https://schema.org","@type":"Article","headline":"PAGE TITLE",
   "datePublished":"2026-01-15","dateModified":"2026-08-01",
   "author":{"@type":"Organization","name":"Custyle"}}
  </script>
Use @type Article for posts/guides, Product for product pages, Organization on the homepage.
Check: paste the page URL into validator.schema.org (0 errors). Then: npx fastergeo verify.

Where: each affected page, in two places — visible text near the title AND machine-readable markup in <head>.
What: visible: an "Updated: <month year>" line under the H1 — use today for pages you touch now, the real publish date otherwise. Machine-readable: add datePublished/dateModified to the page JSON-LD (see the JSON-LD ticket example), or <meta property="article:published_time" content="2026-08-01">.
Check: npx fastergeo verify.

This is a rollup ticket — no separate work. The average rises as you complete the per-issue tickets above (each names its own pages and fix).
Check: after finishing the other page tickets, run npx fastergeo verify; this flips to done when the re-crawl average reaches the target.