greengingerdigital.com — AI Site Grade

Analysis

I have enough to write a sharp audit. Let me compile it.

Green Ginger Digital: AI-Visibility Audit

The site's own homepage publishes a datePublished of January 2026 and a dateModified of May 2026 — dates that do not exist yet — which means every AI crawler ingesting this page is being fed a timestamp from the future, a signal that undermines credibility in any knowledge-graph context.

Crawler Access

The robots.txt is a bare Yoast default with no AI-bot rules — no mention of GPTBot, ClaudeBot, Google-Extended, or any other crawler. In practice, compare_bot_access shows ClaudeBot and Bytespider return 403 (blocked by nginx), while GPTBot, Google-Extended, OAI-SearchBot, and ChatGPT-User all get full 200 responses with identical content to a browser. PerplexityBot and Perplexity-User timed out entirely. The /llms.txt returns a 404 — no AI content map exists. The site runs on a single nginx server (DigitalOcean IP 167.99.81.187) with no CDN or WAF layer, meaning AI crawlers hit the origin directly with no caching protection.

Cold-Knowledge Gap

A frontier LLM queried cold about "Green Ginger Digital" returned no specific knowledge — the model could not confirm services, clients, location, or any achievements. This is striking because the site prominently claims clients including Pret A Manger, MKM Building Supplies, and PACK'D, and boasts Google Premier Partner status, a Management Today Top 35 Women Under 35 award for co-founder Emily Falkingham, and a European Paid Media Awards finalist placement. None of these signals have propagated into LLM training data. The gap between the agency's real-world credentials and what AI models know cold is near-total.

Schema and Content Posture

JSON-LD is present via Yoast but limited to WebPage, WebSite, BreadcrumbList, and Organization types. There is no LocalBusiness schema, no Review schema for the extensive client testimonials, no FAQPage schema despite the site's answer-style content, and no Article schema on case study pages (the PACK'D case study uses only generic WebPage). The homepage has strong visible text (~1,000 words) with clear H1/H2 structure and client logos, but the services page duplicates its service blocks twice in the rendered HTML — a content quality flag. No FAQ, comparison tables, or structured data formats that AI answer engines prefer are present anywhere on the site.

External Signals

Web searches for the brand name combined with major clients (Pret A Manger, MKM, PACK'D) returned zero indexed results on DuckDuckGo. The Wayback Machine holds no snapshots of the domain. The site has no Google Business Profile, no Clutch/G2/Trustpilot profile visible, and no press coverage surfaced in search. The only external footprint is a LinkedIn company page and social links in the footer. For an agency that claims national and international client work, the off-domain citation graph is effectively empty — which explains why LLMs have no prior knowledge to draw on.

Findings

1. Homepage publishes non-existent future dates (January 2026, May 2026) High

   The homepage JSON-LD includes a datePublished of January 2026 and a dateModified of May 2026, which are in the future. This undermines credibility in knowledge graphs and may cause AI crawlers to distrust the page's timeliness.

   What to change: Update the datePublished and dateModified values in the JSON-LD to reflect actual publication and modification dates.

2. No /llms.txt file published Medium

   The site returns a 404 for /llms.txt, meaning there is no AI-specific content map to guide crawlers to key pages or provide structured summaries.

   What to change: Create an /llms.txt file listing important pages (services, case studies, team) with brief descriptions.

3. robots.txt lacks AI-bot directives Medium

   The robots.txt is a bare Yoast default with no rules for GPTBot, ClaudeBot, Google-Extended, or other AI crawlers. While some AI bots are allowed by default, the lack of explicit rules leaves access uncontrolled.

   What to change: Add explicit allow/disallow rules for major AI crawlers (GPTBot, ClaudeBot, Google-Extended, etc.) to the robots.txt file.

4. ClaudeBot and Bytespider are blocked by nginx High

   ClaudeBot and Bytespider receive 403 responses when accessing the site, preventing their content from being indexed by those crawlers.

   What to change: Ensure nginx configuration does not block legitimate AI crawlers; allow ClaudeBot and Bytespider access.

5. PerplexityBot and Perplexity-User time out Medium

   PerplexityBot and Perplexity-User timed out when attempting to access the site, indicating potential server performance or firewall issues.

   What to change: Investigate server response times and firewall rules to ensure Perplexity crawlers can access the site within reasonable timeouts.

6. Missing LocalBusiness schema for physical location Medium

   The site uses Organization schema but lacks LocalBusiness schema, which would help AI models associate the agency with its Hull location and services.

   What to change: Add LocalBusiness schema with address, telephone, and opening hours to the homepage or contact page.

7. No Review schema for client testimonials Medium

   The site displays client testimonials but does not mark them up with Review schema, missing an opportunity to appear in rich snippets and AI answer boxes.

   What to change: Add Review schema to testimonial sections on the homepage and case study pages.

8. Case study pages lack Article schema Medium

   The PACK'd case study page uses only generic WebPage schema instead of Article schema, reducing its chances of being featured in AI-generated summaries.

   What to change: Add Article schema to case study pages with headline, datePublished, and author fields.

9. Services page has duplicated service blocks in HTML Low

   The services page renders its service blocks twice in the HTML, which may be seen as content quality issue by crawlers and could dilute keyword relevance.

   What to change: Remove the duplicated service blocks from the HTML to present clean, unique content.

10. No FAQPage schema on answer-style content Low

    The site has content that answers common questions but does not use FAQPage schema, missing a chance to appear in AI-generated FAQ snippets.

    What to change: Identify pages with question-answer content and add FAQPage schema.

11. Near-zero external citation graph High

    Web searches for the brand name combined with major clients (Pret A Manger, MKM, PACK'D) returned zero indexed results. No Google Business Profile, Clutch, G2, or Trustpilot profiles were found. The Wayback Machine has no snapshots. This explains why LLMs have no prior knowledge of the agency.

    What to change: Build external citations by creating a Google Business Profile, claiming profiles on Clutch/G2/Trustpilot, and securing press coverage or backlinks from reputable sites.

12. LLMs have no knowledge of Green Ginger Digital High

    A frontier LLM queried cold about the agency returned no specific knowledge—could not confirm services, clients, location, or achievements. This is despite the site claiming Pret A Manger, MKM, PACK'D clients, Google Premier Partner status, and awards.

    What to change: Improve external signals (citations, backlinks, press) and ensure schema markup is comprehensive to help LLMs discover and cite the agency.

What's working

• GPTBot and Google-Extended are allowed and return full content — Major AI crawlers like GPTBot, Google-Extended, OAI-SearchBot, and ChatGPT-User receive 200 responses with identical content to a browser, ensuring their content is fully indexed.
• Homepage has strong visible text with clear H1/H2 structure — The homepage contains ~1,000 words of well-structured content with clear headings and client logos, providing good context for AI crawlers.
• Sitemap is present and indexes 80 URLs — The sitemap returns 200 with 80 URLs and is indexed, helping crawlers discover all pages efficiently.
• Organization schema is present on homepage — The homepage includes JSON-LD Organization schema with name, logo, and URL, providing basic entity information to AI crawlers.
• Case study pages contain detailed, narrative content — The PACK'd case study page has over 1,100 words of detailed narrative with results and strategy, providing rich material for AI summarization.
• Team page provides individual bios — The /our-team/ page lists team members with roles and backgrounds, offering personal context that can be used by AI for entity recognition.
• Press release about Pret A Manger client win is published on site — The site hosts a press release announcing Pret A Manger as a client, which is a strong signal for AI crawlers if properly indexed.
• Services page clearly categorizes offerings — The services page lists SEO, PPC, and Paid Social services with descriptions, helping AI crawlers understand the agency's core offerings.