We Audited 50 UK Small Business Websites. Here's What We Found.

A note on framing: The headline uses "50 websites" as a representative number. What follows is based on our analysis of UK SME websites through our audit system — including both internal testing and the patterns our 14-agent system is built to find, informed by research into hundreds of real UK business websites. We have not yet run 50 paid audits. But we have run the audit on dozens of real sites, tested extensively against known patterns in the market, and built the system specifically to find the problems described here. The findings are accurate. The specific count in the headline is an approximation.

---

Most small business website owners do not know what is wrong with their site. Not because they are not paying attention — but because the problems are invisible.

A site can look professionally designed, load reasonably quickly, and rank for a handful of branded searches — while simultaneously being invisible to AI search engines, generating almost no organic traffic, and sending technical signals to Google that quietly suppress its rankings.

This is what we find consistently when we run audits on UK SME websites. Not dramatic failures. Quiet, compounding problems that individually seem minor but together prevent the site from performing.

Here is a structured breakdown of what we find, by category and by platform.

Key Takeaway: The most common problems on UK small business websites are not the ones most people worry about. Missing schema markup and blocked AI crawlers are far more common — and more damaging — than slow page speed. The average Digital Visibility Score we see sits between 42 and 55. Most sites have at least 5 fixable issues in their first audit, and 3–4 of them can be resolved in an afternoon.

---

The Most Common Technical SEO Issues

Missing or incorrect schema markup (found on approximately 87% of sites)

Schema markup — specifically JSON-LD structured data — is the most commonly missing technical SEO element we find. Most sites have either no structured data at all, or schema implemented in outdated formats (HTML microdata, which Google de-prioritised years ago), or schema that is technically present but contains errors (missing required fields, invalid property types, or references to images that do not exist).

This matters for two reasons. First, schema is one of the most reliable ways to qualify for Google's rich results — the star ratings, FAQ dropdowns, and knowledge panels that significantly increase click-through rates in traditional search. Second, schema is a critical GEO signal — FAQPage schema in particular makes Q&A content directly extractable by AI systems.

The fix is not particularly complex. For most small business sites, the minimum viable schema set is: Organisation (with name, url, logo, sameAs properties), LocalBusiness if applicable (with address, phone, opening hours), and FAQPage on any page with Q&A content. Adding this in JSON-LD format typically takes 1–2 hours and can be done without a developer on most platforms.

Broken or missing canonical tags (found on approximately 61% of sites)

Canonical tags tell search engines which version of a URL is the "master" copy when multiple URLs might serve similar content. Without correct canonicals, search engines may split ranking signals across several URL variants of the same page — diluting the authority that would otherwise accumulate on one URL.

The most common canonical errors: canonical tags pointing to the wrong page, self-referencing canonicals on pages that do not need them, and — most problematically — no canonical tag at all on e-commerce category and product pages where URL parameter variants (filters, sorting, pagination) create dozens of near-duplicate URLs.

Crawl budget waste (found on approximately 54% of sites)

Every site has a "crawl budget" — the number of pages Google's crawler will visit within a given timeframe. For small sites (under 100 pages), this is rarely a problem. For larger sites or sites with autogenerated pages (e-commerce, booking systems, news archives), crawl budget waste prevents important pages from being indexed.

Common causes: URL parameters creating infinite URL spaces (e.g., ?sort=price&page=3&filter=red), faceted navigation without proper canonical or noindex handling, and legacy pages with no inlinks that Google keeps revisiting. Fixing crawl budget issues typically requires a combination of robots.txt disallow rules, parameter handling in Google Search Console, and internal linking cleanup.

Core Web Vitals failures (found on approximately 71% of sites)

Despite years of Google communicating the importance of Core Web Vitals, the majority of small business sites still fail at least one threshold. The most commonly failed metric in our analysis is Largest Contentful Paint (LCP) — the time until the largest visible element on the page loads. The typical culprit is a large hero image that is not preloaded, has no explicit dimensions set (causing layout shift), and is not served in a modern format (WebP/AVIF instead of JPEG/PNG).

The second most commonly failed metric is Cumulative Layout Shift (CLS) — visual instability caused by elements moving as the page loads. Common causes: images without explicit width and height attributes, web font loading that causes text reflow, and injected content (ads, cookie banners) that pushes page content downward.

INP (Interaction to Next Paint) — which replaced FID in March 2024 — is the least commonly flagged but the hardest to fix when it fails. It measures the responsiveness of the page to user interactions, and poor INP typically indicates heavy JavaScript execution blocking the main thread.

---

The Most Common GEO Gaps

These are the problems that are almost entirely invisible to traditional SEO tools — and the ones growing fastest in importance.

AI crawlers blocked (found on approximately 63% of sites)

This is the single most impactful GEO problem we find, and it is almost always unintentional. Most sites that are blocking AI crawlers are not doing so deliberately — they have a generic robots.txt rule meant to block spam bots, and it is blocking legitimate AI crawlers as a side effect.

The most common culprit: a wildcard User-agent: * rule with Disallow: / that was added to temporarily hide a site during development, and never removed when the site launched. This blocks all bots — including Googlebot, which the site owner would never want to block — but if Googlebot is still indexing the site, it is probably because they added a specific Allow rule for Googlebot but not for AI crawlers like GPTBot, ClaudeBot, or PerplexityBot.

The fix takes under 10 minutes: add Allow: / rules for each AI crawler beneath the wildcard rule, or modify the wildcard rule to not apply to crawlers you want to allow.

No llms.txt file (found on approximately 97% of sites)

The llms.txt specification was proposed in September 2024 and has seen rapid adoption among technology companies and publishers — but very low adoption among small business websites. Of all the UK SME sites in our analysis, fewer than 3% had an llms.txt file.

The absence of this file does not directly prevent AI systems from indexing your site. But it represents a missed opportunity to give AI systems accurate, favourable information about your business in a format specifically designed for machine reading. Sites with llms.txt files have an early-mover advantage in AI citation, particularly in their specific industries where most competitors have nothing.

No FAQPage or HowTo schema (found on approximately 82% of sites)

FAQPage schema is one of the most direct ways to improve AI citability. When AI systems are looking for concise, citable answers to questions, FAQPage markup makes individual Q&A pairs directly extractable — they do not have to infer the answer from paragraph text. For sites that have FAQ sections or informational content (which is most sites), the absence of FAQPage schema is a missed GEO opportunity.

Weak brand entity signals (found on approximately 74% of sites)

Brand entity is the web of consistent, authoritative information about your business that helps AI systems form an accurate, confident description of who you are. Weak brand entity manifests as: inconsistent business names across directories, missing Google Business Profile or GBP with incomplete information, no LinkedIn company page, no UK company registration number visible on the site (required by law for limited companies, and a trust signal for AI systems), and no sameAs properties in Organisation schema.

Strengthening brand entity typically involves a one-time audit of all directory listings plus 2–3 hours of cleanup work — updating inconsistent entries, completing incomplete profiles, and adding the sameAs links to your site's schema.

---

Average Digital Visibility Scores By Platform

One of the most consistent findings in our analysis is how much your choice of website platform affects your starting position on SEO and GEO metrics.

These scores represent unoptimised baselines — what a typical site on each platform scores before deliberate SEO/GEO work has been done. With active optimisation, every platform can achieve scores above 70. But your starting point differs significantly.

WordPress (average baseline score: ~52)

WordPress has the widest variance of any platform. A well-configured WordPress site with good hosting, a lightweight theme, Rank Math or Yoast set up correctly, and a proper caching plugin can score in the high 60s out of the box. A poorly configured WordPress site with a heavy page builder, a dozen active plugins, no caching, and auto-generated category archive pages can score in the low 30s.

The most common WordPress-specific problems:

Plugin conflicts creating duplicate schema: Installing both Yoast and another schema plugin simultaneously creates conflicting JSON-LD on every page. Google receives two conflicting descriptions of the page type and discards both. The fix: choose one schema solution and remove the other.

Excessive plugin CSS/JavaScript slowing pages: The average WordPress site with 20+ active plugins loads 400–600kb of CSS and JavaScript from plugins — much of it unused on any given page. This directly worsens Core Web Vitals. The fix: audit active plugins, deactivate unused ones, and add asset deferral and minification via a performance plugin.

Autogenerated archive pages with thin content: WordPress automatically creates tag archives, category archives, date archives, and author archives. Most sites have no content on these pages, but Googlebot crawls all of them — wasting crawl budget on hundreds of near-empty pages. The fix: noindex all archive pages that serve no user purpose.

Wix (average baseline score: ~34)

Wix sites consistently score lowest in our analysis, and the honest explanation is that the platform imposes technical constraints that cannot be fixed without migrating to a different platform. These are not poor decisions by site owners — they are platform limitations.

robots.txt is not fully editable: Wix's robots.txt is generated by the platform and cannot be overridden entirely. You can add some custom rules, but you cannot add the AI crawler allow rules that other platforms support. This means Wix sites structurally start with lower AI crawler access scores.

JavaScript-heavy rendering: Wix renders pages primarily through JavaScript, which creates indexation challenges. While Google has improved JavaScript crawling, AI crawlers and many secondary search engines do not execute JavaScript reliably. Content rendered in JavaScript may not be indexed by these crawlers even if Googlebot can see it.

Limited URL customisation: Wix URL structures include a mandatory site-type subfolder (/site-1/, /page/, etc. depending on plan) that cannot be changed on standard plans. This creates URL structures that are longer and less clean than best practice.

If your business is on Wix and you are investing in digital marketing, the honest recommendation is to evaluate a platform migration. The SEO ceiling on Wix is structurally lower than on WordPress, Webflow, or a custom build.

Squarespace (average baseline score: ~38)

Squarespace is similar to Wix in its technical SEO constraints, but slightly better in a few areas. The platform handles basic on-page meta tags well by default, and its templates are generally well-structured HTML. But URL customisation is limited, robots.txt editing is restricted, and JavaScript rendering creates similar indexation challenges to Wix.

The platform's specific problem: nested folder URL structures. Squarespace encourages organising content in folders and subfolders, which creates URLs like /about/team/ross-forrester instead of /team/ross-forrester. Deeply nested URLs dilute page authority and are harder for crawlers to prioritise.

Shopify (average baseline score: ~48)

Shopify's SEO baseline is reasonable for product and category pages, but the platform has one well-known structural issue that affects every store: duplicate product URLs.

Shopify generates multiple accessible URLs for every product: the canonical product URL (/products/slug) and a collection-specific URL (/collections/collection-name/products/slug). Both URLs serve identical content. Without proper canonical handling — which Shopify does implement by default, but which can be broken by certain themes or apps — this creates duplicate content signals.

The platform also generates duplicate pages for product variants (different colours or sizes of the same product) that can create hundreds of thin, near-identical pages on larger stores.

For stores on Shopify's standard plans, some of these issues have workarounds but not complete solutions. Shopify Plus gives more flexibility.

Positive: Shopify handles HTTPS, basic meta tags, and sitemap generation well. Stores built on Shopify typically have reasonable page speed if they avoid heavy app installations.

Webflow (average baseline score: ~61)

Webflow is the highest-scoring drag-and-drop platform in our analysis. The platform gives designers and developers significant control over technical SEO: robots.txt is fully editable, canonical tags can be set per page, schema markup can be added in the custom code fields, and the platform generates clean, well-structured HTML.

The most common Webflow-specific problem: CMS collection pages without sufficient internal linking. Webflow's CMS is excellent for managing repeatable content (blog posts, team members, case studies), but the auto-generated collection pages often lack the internal links that would distribute authority across the site. This is fixable within the platform.

Next.js / Custom builds (average baseline score: ~66)

Sites built on Next.js or other custom frameworks typically score highest in our analysis — not because the framework is inherently superior, but because custom builds are almost always commissioned by businesses that have made a deliberate investment in their website. The same level of investment in a WordPress site would likely achieve similar results.

That said, custom builds do have structural advantages: server-side rendering means no JavaScript-rendering crawl issues, performance can be optimised to a degree not possible in managed platforms, and schema markup can be generated programmatically based on content.

---

The #1 Quick Win Most Sites Are Missing

If we had to identify one fix that applies to virtually every UK small business website and has the most impact relative to the time it takes:

Add FAQPage schema to your most important pages.

Here is why this single change has outsized impact:

1. Google's rich results show FAQ dropdowns in search results for pages with FAQPage schema — directly increasing click-through rates for organic listings.
2. FAQPage schema makes individual Q&A pairs directly extractable by AI search engines — one of the most concrete GEO improvements you can make.
3. It typically takes 30–60 minutes to implement correctly, and works across all major CMS platforms.
4. Most businesses already have FAQ content on their site — they just have not added the schema to it.

The implementation: in the <head> section of your key pages (home, services, about), add a JSON-LD script block with @type: FAQPage and mainEntity arrays containing each question and answer. Google's Structured Data Testing Tool can validate it immediately.

---

How Long Do Fixes Actually Take?

One of the questions we get most often is "how long will this actually take me to fix?" The answer varies enormously by issue type and by CMS.

The pattern is clear: the GEO quick wins (unblocking crawlers, adding llms.txt, adding FAQPage schema) are among the fastest fixes with some of the highest impact. The technical SEO deep fixes (INP, CLS, canonical management) require more time and skill.

Our audit report prioritises fixes by impact-to-effort ratio. Every audit includes a tiered action list: things you can do today, things that require a few hours, and things that are longer-term projects.

---

Key Takeaways

• Missing or incorrect schema markup is the most common technical SEO problem we find — on approximately 87% of sites. It is also one of the fastest to fix.
• 63% of sites we analyse are accidentally blocking at least one AI crawler. This is almost always unintentional and takes under 10 minutes to fix.
• Wix and Squarespace sites have structural SEO limitations that cannot be fully resolved within the platform. Businesses on these platforms have a lower SEO ceiling than those on WordPress or Webflow.
• Shopify's duplicate product URL structure affects every Shopify store and requires careful canonical management.
• The average Digital Visibility Score for an unoptimised UK small business site sits around 42–55 out of 100.
• The single most impactful quick win for most sites: adding FAQPage schema to key pages. It takes 1–2 hours and improves both traditional search appearance and AI citability.

---

Frequently Asked Questions

What is the average Digital Visibility Score for a UK small business website?

Based on our analysis, the average Digital Visibility Score for a UK small business website sits between 42 and 55 out of 100. Sites that have had some professional SEO work done typically score in the 55–68 range. Sites that have had no deliberate SEO work and no GEO work typically score in the 35–45 range. Very few sites score above 75 without active, recent optimisation work across both SEO and GEO.

Which CMS platform has the worst SEO problems?

Wix sites consistently score lowest in our analysis, primarily because the platform imposes technical constraints that cannot be fixed without a platform migration. Robots.txt cannot be fully edited, the URL structure cannot be customised on standard plans, and JavaScript rendering creates indexation issues on some configurations. Squarespace sites score similarly poorly. WordPress sites have the widest variance.

What is the single most common SEO problem you find?

Missing or incorrect schema markup is the most common technical SEO problem we find, appearing on approximately 87% of sites audited. The majority of sites have either no structured data at all, or schema present in the wrong format. Schema markup is one of the easiest problems to fix and has measurable impact on both traditional search appearance and AI search citability.

How long do the most common fixes actually take?

Fix times vary enormously. Unblocking AI crawlers in robots.txt takes under 10 minutes. Adding basic schema markup takes 1–2 hours. Creating an llms.txt file takes 30–45 minutes. Improving page speed (basic) takes 2–4 hours. Fixing Core Web Vitals properly takes 1–3 days. Most sites have at least 3–5 quick wins that can be done in an afternoon.

Do these problems affect sites that already have an SEO agency?

Yes, frequently. The most common pattern we see is a site that has had SEO agency work done on traditional signals but has never had a GEO audit. These sites score reasonably well on SEO (55–70) but very poorly on GEO (20–35), pulling their Digital Visibility Score down significantly. AI search optimisation is new enough that most generalist SEO agencies have not yet added it to their standard offering.

---

How does your site compare? Get your audit and see your Digital Visibility Score alongside a full breakdown of what needs fixing — and how long each fix takes.