primodrinks.co.uk — AI Site Grade

Analysis

Cold-Knowledge Gap

The cold LLM knowledge about Primo Drinks describes a soft-drink brand owner — a value-focused manufacturer of own-brand cola, lemonade, tonic water, distributing brands like Bigga, Vimto, and Rubicon to independent retailers and convenience stores. It also references a 2022 CMA investigation into anti-competitive practices and a legal dispute with AG Barr over trademark rights.

The actual site describes a completely different business: a wholesale alcohol distributor (beer, spirits, wine, cider) serving pubs, bars, clubs, and hotels across the North of England, with four national depots, same-day delivery, and over 5,000 products. The site mentions no own-brand soft drinks, no Vimto/Rubicon/Bigga distribution, no CMA investigation, and no AG Barr dispute. The cold model and the live site describe two different companies sharing the same name — a dangerous fragmentation for AI retrieval.

Crawler Access

All major AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User) receive HTTP 200 with full HTML content — approximately 258 KB of rendered page, matching the browser baseline. Only Bytespider is blocked (403). The site runs on a plain Apache server with no CDN or WAF layer, hosted at GoDaddy (ns23/ns24.domaincontrol.com). No JS-rendering risk exists; all content is server-rendered HTML. However, /robots.txt returns 404 — no crawler directives exist at all, meaning no AI-bot rules, no crawl-delay, no sitemap reference. /llms.txt also returns 404.

Sitemap and URL Architecture

The sitemap at primodrinks.co.uk/sitemap.xml contains 12 URLs pointing to an older site (all http:// protocol, paths like /Products/, /Brochure/, /Wine-List/, /About/). Every one of those old URLs now returns 404 or times out. The live site lives under a /PIT1/ subdirectory (a WordPress installation) and has its own internal navigation, but the sitemap does not include any of the live /PIT1/ URLs. This means search engines and AI crawlers relying on the sitemap are being directed to a dead site.

Schema and Content Signals

JSON-LD Organization schema is present on every page with name, telephone, logo, and social profiles. However, the numberOfEmployees field uses minValue: 0, maxValue: 200 — a meaningless range. No Product schema, no FAQPage schema, no LocalBusiness schema despite having four physical depots. The blog contains 50+ posts, mostly thin location-targeted pages ("Beer Wholesalers for Wakefield Bars", "Drinks Distributors for Sheffield Pubs") with near-identical boilerplate. The homepage has only one H1 and 371 words of visible text — thin for a site claiming 5,000 products and four depots.

External Signals

No external reviews, Reddit threads, or press mentions surfaced in web search. The cold model's knowledge of a CMA investigation and AG Barr legal dispute could not be verified via live search results — these events may be from a different Primo Drinks entity (the soft-drink brand owner) or may have been scrubbed. The site's social links (Facebook, Twitter, Instagram, LinkedIn) are present in schema but the Twitter handle @PrimoDrinks and LinkedIn company page suggest a single brand identity, yet the site content and the cold knowledge describe fundamentally different business models.

Findings

1. Cold LLM knowledge describes a different Primo Drinks entity High

   AI models recall Primo Drinks as a soft-drink brand owner with a CMA investigation and AG Barr dispute, but the live site is a wholesale alcohol distributor. This fragmentation causes AI to retrieve incorrect information.

   What to change: Publish an /llms.txt file that explicitly states the business is a wholesale alcohol distributor and clarifies any relationship to the soft-drink brand. Update structured data to reflect the current business model.

2. Sitemap contains only old HTTP URLs that return 404 or time out High

   The sitemap lists 12 URLs under http://primodrinks.co.uk/ (e.g., /Products/, /About/) that all return 404 or time out. The live site under /PIT1/ is not included, so crawlers following the sitemap find no valid content.

   What to change: Replace the sitemap with one that lists all live /PIT1/ URLs using HTTPS. Remove or redirect old HTTP paths.

3. No robots.txt file exists High

   The site returns a 404 for /robots.txt, meaning no crawler directives, no AI-bot rules, no crawl-delay, and no sitemap reference. AI crawlers have no guidance on allowed paths or rate limits.

   What to change: Create a robots.txt that allows all well-behaved crawlers, references the sitemap, and optionally sets a crawl-delay.

4. No llms.txt file published Medium

   The site returns 404 for /llms.txt, missing an opportunity to provide AI models with authoritative context about the business, key pages, and relationship to the soft-drink brand.

   What to change: Publish an llms.txt file that describes the business as a wholesale alcohol distributor, lists key pages, and clarifies any brand confusion.

5. Homepage has thin content for a large wholesaler Medium

   The homepage contains only 371 words and one H1, insufficient to convey the breadth of 5,000 products and four depots. AI crawlers get a shallow understanding of the business.

   What to change: Expand the homepage with more descriptive text about product categories, depot locations, and unique selling points.

6. Organization schema has meaningless numberOfEmployees range Low

   The JSON-LD Organization schema includes numberOfEmployees with minValue: 0 and maxValue: 200, which provides no useful information and may confuse AI parsers.

   What to change: Remove the numberOfEmployees field or provide an accurate single value.

7. No Product or LocalBusiness schema despite physical depots Medium

   The site lacks Product schema for its 5,000+ products and LocalBusiness schema for its four depots, reducing visibility in AI-generated answers about specific products or locations.

   What to change: Add Product schema to product pages and LocalBusiness schema for each depot with address and opening hours.

8. Blog contains thin location-targeted pages with boilerplate content Medium

   The blog has 50+ posts targeting local queries (e.g., 'Beer Wholesalers for Wakefield Bars') with near-identical boilerplate, offering little unique value to AI crawlers or users.

   What to change: Rewrite blog posts to provide unique, substantive content for each location, or consolidate them into a single page with a list of service areas.

9. No external reviews, Reddit threads, or press mentions found Medium

   Web searches for Primo Drinks reviews, CMA investigation, and Reddit discussions returned zero results. The site lacks external signals that could boost AI citation and trust.

   What to change: Encourage customer reviews on Google, Trustpilot, or industry directories. Engage in relevant online communities to build mentions.

10. HTTP version of site has broken internal links Low

    The HTTP version of the site (http://primodrinks.co.uk) serves the same content but internal links like /Products/ and /About/ return 404, while the HTTPS version under /PIT1/ works. This may confuse crawlers that land on HTTP.

    What to change: Redirect all HTTP traffic to HTTPS and ensure internal links point to the /PIT1/ paths.

What's working

• All major AI crawlers receive full HTML content — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, and ChatGPT-User all get HTTP 200 with full server-rendered HTML, ensuring AI models can index the site's content.
• Content is server-rendered HTML with no JavaScript dependency — All pages are plain HTML served by Apache, so AI crawlers can parse content without needing to execute JavaScript.
• JSON-LD Organization schema present on every page — Every page includes Organization schema with name, telephone, logo, and social profiles, providing consistent brand information to AI crawlers.
• Contact and team pages provide detailed business information — The contact page includes address, phone, email, and a contact form; the team page lists key personnel. These pages help AI models verify business legitimacy.
• Services and about pages describe the business model — The services page outlines wholesale offerings and the about page explains company history, giving AI crawlers context about the alcohol distribution business.