GEO Audit Report

pimlicoplumbers.com

Audited on September 30, 2026

43Needs Work

AI Analysis Summary

This page scores 43/100, indicating a significant need for improvement to enhance AI citability and extractability. Key weaknesses lie in Extractability, On-Page Trust, Schema Markup, and Entity Consistency, which currently hinder AI models from effectively understanding and citing the content. While AI crawler access is strong, the page's content structure, trustworthiness signals, and underlying schema require immediate attention to improve its performance in AI-powered search.

Performance Overview

Category Breakdown

93Excellent
AI Crawler Access
  • All 5 AI crawlers (GPTBot, ChatGPT-User, PerplexityBot, ClaudeBot, Google-Extended) are allowed
  • No llms.txt file found — consider adding one at /llms.txt to expose a curated, AI-friendly site index
10Critical
Entity Consistency
  • No Organization schema detected — add JSON-LD Organization with name, url, logo, and sameAs links so AI engines can resolve the entity
  • Organization schema has no sameAs URLs — list official LinkedIn, X, Crunchbase, Wikipedia, GitHub profiles to disambiguate the entity
  • Brand name varies across surfaces: "Pimlico Plumbers - Property Maintenance Services in London" vs. "Plumbing, Heating & Electrical Services in LondonTrustpilot Custom Widget4.4 out of 5 star rating on TrustpilotTrustpilotTrustpilot Custom Widget4.4 out of 5 star rating on TrustpilotTrustpilotreCAPTCHA" — pick one canonical spelling and use it everywhere
  • Brand "Pimlico Plumbers - Property Maintenance Services in London" does not appear in body content — AI may not associate this page with the entity
65Good
Citability (Structure)
  • Exactly one H1 tag — ideal for AI extraction
  • 8 H2 subheadings provide scannable structure
  • 15 H3 tags add nested structure
  • Heading hierarchy skips levels (e.g. H1 → H3) — AI parsers prefer sequential H1 → H2 → H3
9Critical
Extractability
  • No TL;DR / Key Takeaways block found — add one near the top (3–5 bullets) as a primary AI citation target
  • Only 1/8 H2 sections open with a 40–60 word direct answer — aim for most sections
  • 6/8 section openings start with pronouns or filler — rewrite to stand alone
  • 5 H2 section(s) in the top third but none have a direct, self-contained answer opening — these are your highest-value real estate for AI citations
15Critical
Schema Markup
  • BreadcrumbList schema found and valid
  • WebSite schema found and valid
  • No Organization schema detected — consider adding one
  • No Article schema detected — consider adding one
68Good
Content Authority
  • Author attribution detected — builds trust for AI citations
  • Original data + citable stats appear in the first 20% of the page (18 early stats, 0 first-hand voice markers) — ideal for attribution
  • 251 outbound links but none to recognized authoritative domains
  • 1 inline source attribution detected (Charlie Mullins) — LLMs count these even without hyperlinks
20Critical
On-Page Trust
  • "Trusted by" section found but no recognizable brand names — add alt text or named customer logos
  • Review signal on one platform: Trustpilot — add at least one more
  • No quantified case metrics found — AI engines prefer pages with specific numbers over vague claims
  • No security/compliance certifications referenced (SOC 2, HIPAA, GDPR, ISO 27001…)
68Good
Freshness
  • Structured date found: published 2020-02-24, last updated 2025-12-02
  • Date is in metadata but not visible on the page — render a 'Last updated' string for users and AI alike
  • Content is 301 days old — fresh but consider an update within the next 6 months
  • Recently re-edited — dateModified is more recent than datePublished and within the last year
78Good
Technical SEO
  • Server-Rendered Content: Passed
  • Meta Title: Failed — Title is too long — trim to 60 characters or fewer
  • Meta Description: Failed — Meta description exceeds 155 chars — will be truncated in AI/SERP snippets
  • Canonical Tag: Passed
Server-Rendered Content8358 words rendered server-side
Meta Title212 characters: "Pimlico | Plumbing, Heating & Electrical Services in LondonTrustpilot "

Title is too long — trim to 60 characters or fewer

Meta Description161 characters

Meta description exceeds 155 chars — will be truncated in AI/SERP snippets

Canonical Taghttps://www.pimlicoplumbers.com/
Open Graph Tags3/3 tags present
Viewport Meta Tagwidth=device-width, initial-scale=1, shrink-to-fit=no
HTTPSYes
Response Time1206ms
Content-Type Headertext/html
Language Attributeen
H1 TagPresent
Paragraph Density1.8 avg sentences/paragraph across 85 paragraphs
75Good
Brand Presence (SERP)
  • Strong Reddit presence: 10 mentions found
  • 9 LinkedIn mention(s) — grow professional visibility
  • 8 general web mention(s) found
  • Brand mentioned across 3 platforms — good diversity

Action Plan

Prioritized recommendations to improve your AI SEO readiness

#RecommendationImpactEffortCategory
1

Implement a Concise TL;DR Summary

Add a 3-5 bullet point TL;DR (Too Long; Didn't Read) block immediately after the introductory paragraph to summarize key takeaways, dramatically improving extractability for AI models.

highlowExtractability
1

Add Comprehensive Organization Schema Markup with SameAs Links

Implement a robust Schema.org 'Organization' markup, including 'name', 'logo', 'url', 'contactPoint', and critically, 'sameAs' links to all official social media profiles and other brand presences. This will significantly boost Entity Consistency and overall Schema Markup.

highmediumSchema Markup
1

Display Quantifiable On-Page Trust Signals

Integrate specific trust signals such as named customer testimonials, quantified case metrics (e.g., 'Served 10,000+ homes'), and visible security/compliance certifications. This will address the low On-Page Trust score and build credibility.

highmediumOn-Page Trust
2

Optimize H2 Section Openings for Extractability

Revise H2 section openings to be self-contained (avoiding pronouns or filler words) and ensure they fall within the ideal 40-60 word range. This will make it easier for AI models to extract discrete answers from each section.

highmediumExtractability
2

Integrate Authoritative Hyperlinked Citations

Add hyperlinked citations to external, authoritative domains to support claims and data presented on the page. This is crucial for improving Content Authority, as zero such citations were detected.

highmediumContent Authority
3

Standardize Brand Naming in Metadata and Body Content

Ensure the brand name ('Pimlico Plumbers') is consistently mentioned in the body content. Also, clean up and standardize the `og:site_name` and HTML `<title>` tags to only contain the clear brand name, removing any extraneous widget or rating text.

highlowEntity Consistency
3

Correct Heading Hierarchy Structure

Address instances where heading levels are skipped (e.g., jumping from H1 directly to H3) to maintain a logical and accessible document outline. This improves page structure and citability.

mediumlowCitability
4

Optimize Meta Title and Description Length

Shorten the meta title (currently 212 characters) and meta description (currently 161 characters) to comply with best practices for search engine display. Aim for concise, keyword-rich summaries within recommended character limits.

mediumlowTechnical SEO
Share your results:
https://geo.juma.ai/report/cmuntyr3y000004js0u87vcop

Want to improve your AI visibility?

Get actionable insights and automated workflows to boost your presence in AI-powered search results.