thenewtinsomerset.com — AI Site Grade

Analysis

I have enough data now. Let me write the audit.

The site has zero structured data across every page examined — a luxury hotel estate with 308 sitemap URLs, a podcast, a journal, and an FAQ page, yet not a single JSON-LD schema block was found on any fetched page, including the homepage, hotel, spa, restaurants, FAQ, membership, and sister-estates pages.

Crawler Access

All AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Bytespider, Applebot-Extended — receive a full 200 with identical byte-size content as a browser. No UA-based blocking exists. The site runs on Netlify behind Cloudflare DNS, with a permissive robots.txt (Allow: / for *, only /assets/ disallowed). No AI-specific directives exist. The llms.txt returns a 404 — the site has no AI-friendly content map. The journal listing page (/journal) returns only 50 words of visible text, suggesting JS-dependent rendering that may leave AI crawlers with thin content on blog index pages.

Cold-Knowledge Gap

The LLM knows The Newt as a "luxury hotel, spa, and working farm estate" with Hadspen House, Paradise Garden, Somerset Cider Brandy, and mentions in Condé Nast Traveller and The Telegraph. The site itself confirms Three Michelin Keys and a No.1 Condé Nast Traveller Readers' Choice Award for best UK hotel outside London — but the LLM's cold knowledge does not mention these awards. The site's most distinctive differentiators (Michelin Keys, the Roman Villa Experience, the Beezantium, the Great Garden Escape train from Paddington, the membership model that excludes day tickets) are absent from the model's prior. The site positions membership as the only way to access the gardens (no general day tickets), which is a strong brand moat the LLM does not reflect.

Schema Posture

Zero JSON-LD was found on any page: the homepage, /hotel, /hotel/spa, /restaurants, /faq, /buy-membership, /sister-estates, /journal, /the-newt-podcast. No Hotel, LodgingBusiness, Restaurant, FAQPage, Event, Product, or Organization schema exists. The FAQ page (/faq) has 1,577 words of question-answer content rendered in plain HTML headings — a prime candidate for FAQPage markup — but none is present. The membership page lists pricing tiers (Individual £100, Joint £175, Local £60) with no Product or Offer schema. The site has a podcast with 7 episodes and a journal with seasonal posts, both invisible to knowledge-graph enrichment.

External Signals

The site's own content claims Three Michelin Keys and a Condé Nast Traveller Readers' Choice Award No.1 spot — but web searches for these accolades returned no indexed results from the domain's perspective, suggesting these claims may be recent or not widely cited in external press. The DNS records show heavy tracking infrastructure (Google Tag Manager, HubSpot, Hotjar, Microsoft Clarity, Facebook Pixel, Barracuda email security) but no Google Business Profile or review-platform verification TXT records. The journal page links to a post titled "A Symphony of Colour" dated April 9, 2026 — but the URL /journal/a-symphony-of-colour returns a 404, indicating a broken link between the journal listing and its content.

Findings

1. Zero JSON-LD structured data on any page High

   No JSON-LD schema was found on the homepage, hotel, spa, restaurants, FAQ, membership, sister-estates, journal, or podcast pages. This includes missing Hotel, LodgingBusiness, Restaurant, FAQPage, Product, and Organization markup.

   What to change: Add JSON-LD schema for each page type: Hotel for /hotel, Restaurant for /restaurants, FAQPage for /faq, Product/Offer for /buy-membership, and Organization for the homepage.

2. FAQ page lacks FAQPage schema despite 1,577 words of Q&A content High

   The /faq page contains extensive question-and-answer content rendered in plain HTML headings, but no FAQPage structured data is present, missing an opportunity for rich results in search.

   What to change: Add FAQPage schema with Question/Answer markup to the /faq page.

3. Membership pricing tiers lack Product or Offer schema Medium

   The /buy-membership page lists membership tiers (Individual £100, Joint £175, Local £60) with no structured data, preventing AI systems from understanding pricing and availability.

   What to change: Add Product and Offer schema for each membership tier on /buy-membership.

4. Journal listing links to a 404 page Medium

   The journal listing page links to a post titled 'A Symphony of Colour' (dated April 9, 2026), but the URL /journal/a-symphony-of-colour returns a 404 error, indicating a broken link.

   What to change: Fix the broken link or remove the reference to the missing journal post.

5. Journal listing page has only 50 words of visible text Medium

   The /journal page returns only 50 words of visible text, suggesting JavaScript-dependent rendering that may leave AI crawlers with thin content on the blog index.

   What to change: Ensure the journal listing page renders meaningful content server-side or pre-renders for crawlers.

6. No llms.txt file for AI-friendly content map Medium

   The site returns a 404 for /llms.txt, meaning there is no AI-friendly content map to guide language models to key pages.

   What to change: Create an llms.txt file listing key pages (hotel, restaurants, spa, FAQ, membership, podcast) to guide AI crawlers.

7. Top awards (Three Michelin Keys, Condé Nast Traveller No.1) absent from LLM cold knowledge High

   The site prominently claims Three Michelin Keys and a Condé Nast Traveller Readers' Choice Award for best UK hotel outside London, but these accolades are not reflected in the LLM's prior knowledge, reducing AI visibility.

   What to change: Ensure these awards are cited on high-authority external sites and include structured data (e.g., Award schema) on the site.

8. Unique differentiators (Roman Villa, Beezantium, membership model) absent from LLM knowledge Medium

   The site's distinctive features—Roman Villa Experience, Beezantium, Great Garden Escape train, membership-only garden access—are not present in the LLM's cold knowledge, limiting AI-driven discovery.

   What to change: Create dedicated pages for these unique features and ensure they are well-linked and indexed; consider adding schema markup.

9. No Google Business Profile or review-platform verification TXT records Low

   DNS TXT records show heavy tracking infrastructure but no verification records for Google Business Profile or major review platforms, which may affect local search visibility.

   What to change: Add Google Business Profile verification TXT record and consider review platform verification records.

What's working

• All AI crawlers allowed with full content access — All tested AI crawlers (GPTBot, ClaudeBot, PerplexityBot, etc.) receive a 200 response with identical content as a browser, with no UA-based blocking.
• Permissive robots.txt allows all crawlers — The robots.txt allows all user agents with Allow: /, only disallowing /assets/, and no AI-specific restrictions.
• Sitemap available with 80 URLs — A sitemap is served at /sitemap.xml with 80 URLs, helping crawlers discover content.
• Key pages have substantial text content — Pages like /faq (1,577 words), /plan-your-visit (1,232 words), and /the-newt-podcast (881 words) provide rich content for AI crawlers to index.
• Podcast and journal provide ongoing content — The site features a podcast with 7 episodes and a journal with seasonal posts, offering fresh content for AI indexing.
• Membership model clearly communicated — The /buy-membership page clearly explains the membership-only access model with pricing tiers, which is a strong differentiator.