peek-cloppenburg.de — AI Site Grade

Analysis

I have all the data I need. Let me write the audit.

Peek & Cloppenburg (peek-cloppenburg.de) — AI-Visibility Audit

Every major AI crawler — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, Bytespider, Applebot-Extended, anthropic-ai — receives a 403 block from Cloudflare on the primary e-commerce domain, while browser requests return a full 200 with ~634 KB of content. The shop is invisible to the AI ecosystem.

Crawler Access

The robots.txt at peek-cloppenburg.de contains zero AI-bot directives — no mention of GPTBot, ClaudeBot, or any other crawler. The wildcard rule (User-agent: *) allows / broadly, but Cloudflare's WAF overrides this at the edge, blocking all AI user-agents before they reach the application. The corporate site at peek-cloppenburg.com has a permissive robots.txt (Allow: /) and all AI bots pass through with 200 — but this site is a thin brand hub (~194 words of text) with no product data, no sitemap, and no e-commerce functionality. The llms.txt returns a 404 error page on both domains. The sitemap referenced in robots.txt (/sitemaps/de/spider-sitemapindex.xml.gz) exists but is gzip-compressed binary, and the uncompressed path returns 404 — meaning standard XML sitemap discovery is broken for crawlers that cannot decompress .gz inline.

Cold-Knowledge Gap

A frontier LLM queried cold about Peek & Cloppenburg correctly identifies it as a German fashion retailer founded 1901 in Dusseldorf with mid-to-premium brands and over 100 stores. It also knows the critical structural fact: two separate, unrelated companies share the name Peek & Cloppenburg (Dusseldorf vs. Hamburg). The site itself acknowledges this in fine print ("Es gibt zwei unabhangige Unternehmen..."). However, the LLM's prior knowledge is richer on the brand's history and corporate structure than on the actual online shop — it mentions "tailored suits, formalwear" and "private-label collections" but has no awareness of the 62-day return policy, the INSIDER loyalty program, the STYLEBOP high-fashion curation, or the 70+ German store locations. The site's actual value propositions (free returns, 1-2 day delivery, 300+ brands, click-and-collect) are absent from the model's cold recall.

Schema Posture

The homepage carries Organization, WebPage, and WebSite schema with correct address, phone, and social links — but no Product schema, no ItemList schema, no Offer schema on any page examined. The category page (/de/damen/bekleidung/kleider) has only a BreadcrumbList schema despite listing 6,396 products with prices. No FAQ schema, no ProductGroup, no AggregateOffer. The Organization schema names the legal entity as "Fashion ID GmbH & Co. KG" — not "Peek & Cloppenburg" — which creates a mismatch between the brand name consumers search for and the entity AI engines index.

External Signals

The DNS records reveal an anthropic-domain-verification TXT token, confirming the brand has already engaged with Anthropic for some verification purpose — yet ClaudeBot is still blocked at the Cloudflare edge. The site runs on Cloudflare with Kameleoon A/B testing, Baqend speedkit, and OneTrust consent management. The corporate .com domain is fully open to AI crawlers but contains no product inventory, no pricing, and no shopping functionality — it is a static brand brochure. The actual e-commerce engine (Next.js with SSR, served via Cloudflare) renders full HTML to browsers but the WAF configuration treats all known AI user-agents as threats.

Findings

1. All major AI crawlers blocked by Cloudflare WAF on e-commerce domain High

   Every tested AI crawler (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, Bytespider, Applebot-Extended, anthropic-ai) receives a 403 response from Cloudflare on peek-cloppenburg.de, while browser requests return 200 with full content.

   What to change: Reconfigure Cloudflare WAF to allow known AI crawler user-agents, or create a bypass rule for verified bots.

2. robots.txt lacks any AI-bot directives High

   The robots.txt file at peek-cloppenburg.de contains no directives for GPTBot, ClaudeBot, or any other AI crawler. The wildcard rule allows all user-agents, but Cloudflare blocks AI bots regardless.

   What to change: Add explicit Allow directives for AI crawlers in robots.txt to signal intent, and ensure Cloudflare respects them.

3. llms.txt returns 404 on both domains Medium

   The llms.txt file, which provides AI crawlers with a structured overview of the site, is not present on either peek-cloppenburg.de or peek-cloppenburg.com, returning a 404 error.

   What to change: Create an llms.txt file summarizing key pages, product categories, and store locations for AI crawlers.

4. Sitemap is gzip-compressed and uncompressed path returns 404 High

   The sitemap referenced in robots.txt exists only as a gzip-compressed file at /sitemaps/de/spider-sitemapindex.xml.gz. The uncompressed .xml path returns 404, breaking standard XML sitemap discovery for crawlers that cannot decompress .gz inline.

   What to change: Serve the sitemap as uncompressed XML or provide a Content-Type header that allows crawlers to handle gzip. Also serve the uncompressed version at the .xml path.

5. No Product, ItemList, or Offer schema on category or product pages High

   The category page for dresses lists 6,396 products with prices but only contains BreadcrumbList schema. No Product, ItemList, Offer, or AggregateOffer schema is present, depriving AI crawlers of structured product data.

   What to change: Add Product, ItemList, and Offer schema markup to category and product pages, including prices, availability, and product names.

6. Organization schema uses legal entity name instead of brand name Medium

   The Organization schema on the homepage names the legal entity 'Fashion ID GmbH & Co. KG' rather than 'Peek & Cloppenburg', creating a mismatch between the brand consumers search for and the entity AI engines index.

   What to change: Add an alternateName property with 'Peek & Cloppenburg' or use the brand name as the primary name in Organization schema.

7. LLM cold knowledge lacks awareness of key online shop features Medium

   A frontier LLM correctly identifies the brand but has no knowledge of the 62-day return policy, INSIDER loyalty program, STYLEBOP high-fashion curation, 70+ store locations, free returns, 1-2 day delivery, 300+ brands, or click-and-collect service.

   What to change: Publish structured data and content that explicitly describes these value propositions, and ensure AI crawlers can access the pages that mention them.

8. Corporate site is open to AI crawlers but contains no product data Medium

   The peek-cloppenburg.com domain allows all AI bots with 200 responses, but it is a thin brand hub with only ~194 words of text, no product inventory, no pricing, and no shopping functionality. It provides no value for AI visibility of the online shop.

   What to change: Either redirect AI crawlers from the corporate site to the e-commerce domain, or add product summaries and links to the corporate site.

9. Anthropic domain verification token exists but ClaudeBot is still blocked Medium

   DNS records include an anthropic-domain-verification TXT token, indicating prior engagement with Anthropic, yet ClaudeBot receives a 403 from Cloudflare on the e-commerce domain.

   What to change: Ensure that the Cloudflare WAF allows ClaudeBot, given the verification token already in place.

10. No FAQ schema on pages with common questions Low

    Pages that likely contain frequently asked questions (e.g., about returns, delivery) do not use FAQ schema, missing an opportunity to appear in AI-generated answers.

    What to change: Add FAQ schema to pages that answer common customer questions.

What's working

• Homepage includes Organization, WebPage, and WebSite schema — The homepage contains structured data with Organization, WebPage, and WebSite schema, including correct address, phone, and social links, which helps AI systems identify the business entity.
• Corporate site allows all AI crawlers with 200 responses — The peek-cloppenburg.com domain returns 200 for all tested AI bots, with a permissive robots.txt, ensuring at least the brand's corporate presence is visible to AI.
• LLM cold knowledge correctly identifies brand history and structure — A frontier LLM correctly identifies Peek & Cloppenburg as a German fashion retailer founded in 1901 in Düsseldorf, with mid-to-premium brands and over 100 stores, and knows about the two separate companies sharing the name.
• Anthropic domain verification token is configured in DNS — The DNS records include an anthropic-domain-verification TXT token, indicating proactive engagement with Anthropic for verification purposes.
• Category page includes BreadcrumbList schema — The dresses category page uses BreadcrumbList schema, helping AI crawlers understand the site's navigation hierarchy.