Technical SEO checklist for UK marketing sites

Open Search Console on a Monday morning in Leeds and you will often see the same pattern: impressions holding or even rising while organic demo requests and form fills stall. The content calendar is full. The paid team is still burning budget on the same head terms. What quietly failed is not “SEO strategy” in the abstract. It is crawl access, index hygiene, template speed, and the boring wiring between page types and conversion events. That is the channel job for a UK marketing site: earn the click, survive the SERP, and hand a clean session to a page that can convert without thrashing Core Web Vitals.

Another tactics list will not fix that. A scored technical audit will. You mark each item pass, weak, or fail against evidence you can pull this week, then you fix in an order that stops waste before you scale content or bid more aggressively on brand. The latest CrUX data put full three-metric passes at roughly 55-56% of origins (update month on publish); plenty of marketing sites that “look fine” in a desktop walkthrough still sit on the wrong side of that line on mobile, where most discovery traffic lands.

This checklist is written for CMOs, heads of growth, and founders who own the marketing number in GBP, not for hobbyist SEO tinkering. Score the site. Rank the leaks. Spend the next 30 days repairing what actually blocks qualified organic pipeline.

Score every template your UK marketing site ships to Google

Work through the items below with Search Console, CrUX field data, a crawl export, and one live mobile session on your highest-intent landing templates. Do not score from memory. Do not give yourself a pass because the homepage looks sharp on a fibre connection in London.

Audit scorecard

• 1 Crawl budget and robots discipline Inspect robots.txt, parameter URLs, faceted filters, staging hosts, and PDF dumps. Good looks like a short allowlist of public marketing paths, no accidental Disallow on /blog/ or /resources/, and staging locked behind auth rather than “security through obscurity”. Fail costs: Googlebot burns cycles on junk while new service and comparison pages wait days to appear. Open the robots.txt checker and confirm what you think is blocked actually is.

Inspect robots.txt, parameter URLs, faceted filters, staging hosts, and PDF dumps. Good looks like a short allowlist of public marketing paths, no accidental Disallow on /blog/ or /resources/, and staging locked behind auth rather than “security through obscurity”. Fail costs: Googlebot burns cycles on junk while new service and comparison pages wait days to appear. Open the robots.txt checker and confirm what you think is blocked actually is.

• 2 XML sitemap honesty Your sitemap should list every canonical URL you want indexed - money pages, intent clusters, and cornerstone guides - and nothing else. No thank-you pages, no UTM copies, no expired campaign microsites. Good looks like lastmod that moves when content really changes, and Search Console “Sitemaps” showing submitted ≈ indexed within a sensible band. Fail costs: soft-404 noise dilutes discovery of the pages that should drive MQL volume. Run a sitemap check before you argue about content quality.

Your sitemap should list every canonical URL you want indexed - money pages, intent clusters, and cornerstone guides - and nothing else. No thank-you pages, no UTM copies, no expired campaign microsites. Good looks like lastmod that moves when content really changes, and Search Console “Sitemaps” showing submitted ≈ indexed within a sensible band. Fail costs: soft-404 noise dilutes discovery of the pages that should drive MQL volume. Run a sitemap check before you argue about content quality.

• 3 Index coverage without vanity bloat In Coverage / Pages, separate “Excluded by design” from “Should rank but cannot”. Thin tag archives, printer views, and near-duplicate location stubs common on multi-city UK brands (Manchester, Birmingham, Bristol clones with one paragraph swapped) belong noindexed or consolidated. Good looks like every primary commercial URL indexed once. Fail costs: internal competition and diluted signals so none of the variants rank for the commercial query you actually care about. Spot-check with an index status tool on the ten URLs that should produce pipeline.

In Coverage / Pages, separate “Excluded by design” from “Should rank but cannot”. Thin tag archives, printer views, and near-duplicate location stubs common on multi-city UK brands (Manchester, Birmingham, Bristol clones with one paragraph swapped) belong noindexed or consolidated. Good looks like every primary commercial URL indexed once. Fail costs: internal competition and diluted signals so none of the variants rank for the commercial query you actually care about. Spot-check with an index status tool on the ten URLs that should produce pipeline.

• 4 Canonical and hreflang sanity Self-referencing canonicals on live templates; cross-domain or cross-host canonicals only when intentional. If you run .com plus .co.uk or language variants, hreflang must be reciprocal and consistent with the canonical you want UK searchers to land on. Good looks like Search Console International Targeting (or equivalent URL inspection) without conflict errors. Fail costs: the wrong regional template wins, or neither does. Validate canonicals on product, pricing, and lead-magnet templates, not only the homepage.

Self-referencing canonicals on live templates; cross-domain or cross-host canonicals only when intentional. If you run .com plus .co.uk or language variants, hreflang must be reciprocal and consistent with the canonical you want UK searchers to land on. Good looks like Search Console International Targeting (or equivalent URL inspection) without conflict errors. Fail costs: the wrong regional template wins, or neither does. Validate canonicals on product, pricing, and lead-magnet templates, not only the homepage.

• 5 HTTPS, mixed content, and host normalisation One preferred host (www or bare), forced HTTPS, HSTS where appropriate, zero mixed-content console errors on key templates. Marketing sites still ship http asset calls from old CMSs and tag managers. Good looks like clean padlock, single redirect hop from http and non-preferred host. Fail costs: trust friction, broken tracking, and avoidable redirect chains that punish both LCP and bots.

One preferred host (www or bare), forced HTTPS, HSTS where appropriate, zero mixed-content console errors on key templates. Marketing sites still ship http asset calls from old CMSs and tag managers. Good looks like clean padlock, single redirect hop from http and non-preferred host. Fail costs: trust friction, broken tracking, and avoidable redirect chains that punish both LCP and bots.

• 6 Core Web Vitals on real UK mobile traffic Use field data (CrUX / Search Console Experience), not a single lab run on the office Mac. LCP, INP, and CLS must clear thresholds on the templates that carry paid and organic landing traffic. The latest CrUX data put full three-metric passes at roughly 55-56% of origins (update month on publish); if your origin fails, assume mobile conversion rate is already paying a quiet tax. Good looks like stable greens on money URLs for 28-day windows. Fail costs: ranking drag plus form abandon before the CTA paints. Pair a Core Web Vitals check with one filmed mobile scroll on 4G.

Use field data (CrUX / Search Console Experience), not a single lab run on the office Mac. LCP, INP, and CLS must clear thresholds on the templates that carry paid and organic landing traffic. The latest CrUX data put full three-metric passes at roughly 55-56% of origins (update month on publish); if your origin fails, assume mobile conversion rate is already paying a quiet tax. Good looks like stable greens on money URLs for 28-day windows. Fail costs: ranking drag plus form abandon before the CTA paints. Pair a Core Web Vitals check with one filmed mobile scroll on 4G.

• 7 Title, H1, and meta description integrity One H1 that matches the commercial job of the page. Titles unique, under truncation risk, with the primary UK query near the front where it still reads as English. Meta descriptions written for SERP CTR, not stuffed. Good looks like no duplicate titles across service lines, no “Home” or “Untitled” leftovers after a CMS migration. Fail costs: you win impressions and lose the click to a sharper competitor snippet. Spot-check with title, H1, and meta description checkers on the top twenty landing URLs by clicks.

One H1 that matches the commercial job of the page. Titles unique, under truncation risk, with the primary UK query near the front where it still reads as English. Meta descriptions written for SERP CTR, not stuffed. Good looks like no duplicate titles across service lines, no “Home” or “Untitled” leftovers after a CMS migration. Fail costs: you win impressions and lose the click to a sharper competitor snippet. Spot-check with title, H1, and meta description checkers on the top twenty landing URLs by clicks.

• 8 Internal link graph that ships equity to money pages Cornerstone guides and blog posts must pass links, with descriptive anchors, into service, pricing, and comparison URLs. Orphan landing pages created for campaigns are a classic UK agency-site failure. Good looks like every commercial URL reachable within a few clicks from the homepage and from related content, with no “footer only” dependency. Fail costs: pages that rank weakly forever despite decent copy. Map orphans with an internal link analyser before commissioning another 2,000-word post.

Cornerstone guides and blog posts must pass links, with descriptive anchors, into service, pricing, and comparison URLs. Orphan landing pages created for campaigns are a classic UK agency-site failure. Good looks like every commercial URL reachable within a few clicks from the homepage and from related content, with no “footer only” dependency. Fail costs: pages that rank weakly forever despite decent copy. Map orphans with an internal link analyser before commissioning another 2,000-word post.

• 9 Structured data that matches visible proof Organisation, WebSite/Sitelinks searchbox where relevant, Article or FAQ only when the visible page earns it, Product/Offer only if you truly sell that way. No fake review stars. For most UK marketing sites the highest-return types are BreadcrumbList (improves SERP path display), FAQPage on genuinely question-driven pages, and SitelinksSearchbox if branded search volume justifies it. Product and Review markup carry manual-action risk if the page does not match. Good looks like rich-result eligibility without manual actions, and FAQ markup that mirrors on-page copy. Fail costs: ignored markup or, worse, spammy structured data that invites scrutiny. Validate with a schema markup tool and a FAQ schema generator only after the copy is final.

Organisation, WebSite/Sitelinks searchbox where relevant, Article or FAQ only when the visible page earns it, Product/Offer only if you truly sell that way. No fake review stars. For most UK marketing sites the highest-return types are BreadcrumbList (improves SERP path display), FAQPage on genuinely question-driven pages, and SitelinksSearchbox if branded search volume justifies it. Product and Review markup carry manual-action risk if the page does not match. Good looks like rich-result eligibility without manual actions, and FAQ markup that mirrors on-page copy. Fail costs: ignored markup or, worse, spammy structured data that invites scrutiny. Validate with a schema markup tool and a FAQ schema generator only after the copy is final.

• 10 JS rendering and critical content in the initial response Primary copy, H1, pricing cues, and primary CTA must not depend on a late client-side render that Search Console’s rendered HTML never sees. Marketing sites on heavy frameworks fail this more than brochure WordPress builds. Good looks like URL Inspection “rendered” view showing the same commercial claims users see. Fail costs: rankings for thin shells while humans see a full page, or the reverse in edge cases. Compare view-source versus rendered HTML on two high-intent URLs.

Primary copy, H1, pricing cues, and primary CTA must not depend on a late client-side render that Search Console’s rendered HTML never sees. Marketing sites on heavy frameworks fail this more than brochure WordPress builds. Good looks like URL Inspection “rendered” view showing the same commercial claims users see. Fail costs: rankings for thin shells while humans see a full page, or the reverse in edge cases. Compare view-source versus rendered HTML on two high-intent URLs.

• 11 Pagination, parameters, and faceted paths Resource hubs, customer story libraries, and blog indexes need clean pagination or “load more” patterns that do not spawn infinite parameter indexes. Sort and filter URLs should be noindex/follow or canonicalised to a root as policy, not left to chance. Good looks like a written rule your CMS and marketing ops both follow. Fail costs: thousands of near-duplicates soaking crawl time after a redesign.

Resource hubs, customer story libraries, and blog indexes need clean pagination or “load more” patterns that do not spawn infinite parameter indexes. Sort and filter URLs should be noindex/follow or canonicalised to a root as policy, not left to chance. Good looks like a written rule your CMS and marketing ops both follow. Fail costs: thousands of near-duplicates soaking crawl time after a redesign.

• 12 Status codes, soft 404s, and redirect debt Crawl for chains longer than one hop, loops, 404s on linked nav items, and soft 404s that return 200 with “not found” copy. UK rebrands and microsite sunsets leave long tails of broken proof links. Good looks like a quarterly redirect map owned by marketing ops, not a spreadsheet lost in Drive. Fail costs: link equity evaporation and trust hits on the exact URLs PR and partners still cite.

Crawl for chains longer than one hop, loops, 404s on linked nav items, and soft 404s that return 200 with “not found” copy. UK rebrands and microsite sunsets leave long tails of broken proof links. Good looks like a quarterly redirect map owned by marketing ops, not a spreadsheet lost in Drive. Fail costs: link equity evaporation and trust hits on the exact URLs PR and partners still cite.

• 13 Conversion measurement on organic landings GA4 + GTM events for primary convert (demo, trial, call, chat), Enhanced conversions where appropriate, and UTMs that do not create indexable clones. Technical SEO without event integrity leaves you optimising for sessions while pipeline attribution lies. Good looks like the same key events firing on mobile and desktop templates, tested after each tag release. Fail costs: you cannot prove organic contribution, so technical fixes lose budget fights to channels with flashier dashboards. Build test links with a UTM builder; never publish those URLs into the sitemap.

GA4 + GTM events for primary convert (demo, trial, call, chat), Enhanced conversions where appropriate, and UTMs that do not create indexable clones. Technical SEO without event integrity leaves you optimising for sessions while pipeline attribution lies. Good looks like the same key events firing on mobile and desktop templates, tested after each tag release. Fail costs: you cannot prove organic contribution, so technical fixes lose budget fights to channels with flashier dashboards. Build test links with a UTM builder; never publish those URLs into the sitemap.

• 14 Content systems that do not poison technical health Publishing cadence is a technical problem when every post spawns thin author archives, duplicate tag pages, uncompressed hero images, and unlinked drafts on staging. With AI-assisted drafts common across the SERP, the differentiator is governed shipping: human edit, unique titles, internal links, and image weight budgets. Good looks like a release checklist your writers cannot skip. Fail costs: more URLs, worse averages, softer rankings. If volume is the bottleneck, a managed cadence such as HeyLead Auto Blogger can keep UK- and niche-aware articles moving under human strategy without flooding the index with half-finished templates.

Publishing cadence is a technical problem when every post spawns thin author archives, duplicate tag pages, uncompressed hero images, and unlinked drafts on staging. With AI-assisted drafts common across the SERP, the differentiator is governed shipping: human edit, unique titles, internal links, and image weight budgets. Good looks like a release checklist your writers cannot skip. Fail costs: more URLs, worse averages, softer rankings. If volume is the bottleneck, a managed cadence such as HeyLead Auto Blogger can keep UK- and niche-aware articles moving under human strategy without flooding the index with half-finished templates.

DIY tools for this scorecard

• Robots.txt checker

• Sitemap checker

• Google index checker

• Core Web Vitals checker

• Canonical checker

• Internal link analyser

If crawl access and index hygiene already look broken, pause new content sprints until the foundation is scored. For teams that want a specialist to own ongoing technical SEO rather than another one-off audit deck, HeyLead’s technical SEO service is built around this kind of continuous repair, not a single crawl PDF.

Mark pass, weak, or fail, then order the repairs

Keep scoring binary enough that two people would land in the same bucket. Pass means evidence in the last 28 days. Weak means partial coverage, one broken template family, or lab-only optimism. Fail means systematic breakage or no evidence at all.

Weight fails by pipeline impact, not by SEO theatre. A broken canonical on the pricing template outranks a missing FAQ schema on a thought-leadership post. Tracking that drops organic convert events outranks a slightly long title tag. INP failures on the demo form outrank a polite robots comment tidy-up.

Fix order that usually holds for UK marketing sites:

• Access and correctness first: robots, host, HTTPS, status codes, sitemaps.

• Index intent second: canonicals, duplicates, parameters, soft 404s.

• Experience third: CWV on money templates, render-critical content, image weight.

• On-page SERP fourth: titles, H1s, metas, internal links, eligible schema.

• Measurement fifth in parallel once tags stop lying: GA4 events, consent mode behaviour, Enhanced conversions.

• Scale content only after the graph and templates can absorb new URLs without creating orphans.

Resist the urge to “do CWV” as a branding project while Search Console still shows the commercial cluster excluded. Also resist rewriting every blog post before you consolidate five near-duplicate service URLs fighting for the same query in England and Scotland SERPs.

When weak items cluster on landing templates that also take paid traffic, treat them as shared revenue leaks. Paid CPC pressure is not an excuse to ignore the page; it is a reason to fix the template once for both channels. A short soft check: if your team cannot explain which single URL should rank for each primary commercial query, you do not have an information architecture problem hiding as a content problem.

Thirty days of repairs after the scorecard, not a new strategy deck

Use the calendar as a repair sequence. You are not launching a rebrand. You are removing friction Google and users already trip on.

30-day fix playbook

• 1Days 1-3: Export Search Console Pages + Experience, crawl the public site, screenshot robots and sitemap. List every Fail with the URL pattern and owner (dev, web, marketing ops).

• 2Days 4-7: Ship host/HTTPS/robots/sitemap corrections. Remove staging from the public web. Submit updated sitemap. Re-test index status on the ten commercial URLs.

• 3Days 8-12: Collapse duplicates: canonicals, parameter rules, near-duplicate location or campaign pages. 301 where equity matters; noindex where the URL must exist for UX only.

• 4Days 13-18: Attack CWV on the top five landers by organic + paid sessions. Compress heroes, defer non-critical scripts, fix CLS from late fonts and injected banners, reduce main-thread work that wrecks INP on forms.

• 5Days 19-23: Titles, H1s, metas, and internal links for money pages and their supporting guides. Add only schema that mirrors visible content. Validate.

• 6Days 24-27: QA GTM/GA4 convert events on mobile Safari and Chrome. Align thank-you and event names with how you report pipeline. Kill indexable UTM copies.

• 7Days 28-30: Re-score the card. Document residual Weaks with dates. Only then unlock a controlled content burst, with every new URL forced through the release checklist (unique title, internal links, image budget, no new thin taxonomies).

Two realistic patterns show up repeatedly. A B2B SaaS site in London ran four parallel “solutions” URLs with overlapping copy; consolidating to one canonical per job-to-be-done and fixing internal links lifted the surviving URLs without a single new article. A multi-office professional services brand kept a 14-hop redirect chain from an old .co.uk microsite into the main services hub; cutting to one hop and repairing CWV on that hub did more for qualified organic enquiries than another quarter of unfocused blogging.

If post-click behaviour still looks wrong after technical passes, watch session-level friction on forms and CTA scroll depth with something purpose-built for evidence, not opinions. HeyLead Insights is useful when you need to see where UK mobile users stall after the organic click.

Need depth beyond an in-house sprint? Book a focused technical pass via HeyLead SEO technical work once your scorecard shows which template families are failing, so the engagement starts on evidence rather than guesswork.

Two patterns we see repeatedly

“We kept commissioning guides while Search Console showed the pricing templates as crawled, not indexed. The fix was canonical and internal links, not another thought piece. Demo requests from organic moved only after that hygiene landed.” - James, Head of Growth at a Series B HR-tech platform

“INP on the contact form was the silent killer. Lab Lighthouse looked acceptable on office Wi-Fi. Field data on UK mobile was a different story, and paid plus organic both suffered until scripts were cut.” - Sarah, VP Marketing at a multi-office professional services firm

FAQs

How often should a UK marketing site re-run this technical SEO checklist?

Full scorecard quarterly, or after any CMS, design system, tag manager, or domain change. Spot-check money templates monthly in Search Console Experience and Pages. Waiting for a traffic cliff is how technical debt becomes a board conversation.

Is technical SEO still worth it when AI Overviews and zero-click results suppress CTR?

Yes, because the pages that still earn the click and the citation need to be crawlable, fast, and unambiguous. Zero-click pressure raises the cost of every wasted impression; it does not make broken canonicals irrelevant. Pair technical health with sharper intent pages and clearer offers rather than abandoning organic.

Should we prioritise Core Web Vitals or index coverage first?

Index coverage and access first if commercial URLs are excluded, canonicalised away, or trapped behind robots mistakes. CWV next on the URLs that already receive impressions and sessions. Polishing speed on pages Google will not index is theatre.

What belongs in-house versus with a specialist partner?

In-house can own content briefs, offer clarity, and weekly Search Console triage. Specialist capacity usually pays for crawl architecture, render debugging, large-scale redirects, CWV engineering across design systems, and keeping measurement honest when marketing ships fast. The checklist tells you which bucket each fail sits in.

Do we need new content while fixing technical issues?

Pause net-new URL volume if fails cluster on crawl, duplicates, or orphans. Keep refreshing proof and stats on existing money pages. Resume cadence only with a release gate so publishing does not reintroduce thin taxonomies and overweight assets.

Putting it to work

This week

• Score all 14 items with written evidence links (GSC screenshots, crawl rows, CrUX).

• Pick the three Fails that touch commercial URLs or convert events.

• Ship robots/sitemap/host corrections if any of those failed.

• Run Core Web Vitals and index checks on your top ten landers.

• File one engineering ticket per template family, not a vague “SEO improvements” epic.

• Re-test one primary convert event on mobile after any tag change.

Next 30 days

• Complete the repair playbook in order: access, index intent, CWV, on-page, measurement.

• Collapse duplicate commercial URLs and fix internal links into the survivors.

• Install a non-negotiable publish checklist before the next content burst.

• Re-score and keep Weaks visible on the marketing ops board until closed.

• Only then scale programmatic or AI-assisted drafting under human edit and entity-aware internal linking.

Pull your last 28 days of Search Console Page and Experience reports side by side with the fourteen-row scorecard, and refuse any new content budget until the Fail column on commercial templates is empty or ticketed with dates. When the work you need owned is the unglamorous loop of crawl access, index hygiene, template speed, and clean organic conversion wiring on a UK marketing site, HeyLead can run that technical SEO programme as an ongoing partnership so your team stays on offer, narrative, and pipeline judgment. Start the conversation at [email protected].