A technical SEO checklist for Cardiff media and public bodies

What to take away

What Cardiff media and public bodies have in common technically

Cardiff is the Welsh capital and its web estate splits into two big camps. Media sites publish fast, in volume, and often in two languages. Public-sector sites publish slowly, under procurement rules, and also in two languages.

By name, that estate includes BBC Cymru Wales and ITV Cymru Wales, both with bases in the city, WalesOnline as the largest commercial news publisher, and Cardiff University as a bilingual publisher of research, course and policy pages. On the public side it includes Cardiff Council, the Welsh Government and the Senedd Commission, whose own sites and documents are cited across the city's media.

Both run on templates controlled by a central digital team. Both carry archives that stretch back years. Both answer to a Welsh audience that expects to switch language mid-task without losing their place.

The practical overlap is template governance. A single header component, a single cookie banner, a single search results page. Fix the template and you fix hundreds of URLs at once.

The divergence is risk. A media site can ship a broken canonical on Friday and fix it on Monday. A public body cannot, because the change may need an accessibility review and a record of the decision.

That is why a shared technical SEO quality checklist helps both camps: it forces the same evidence for a template change, whether the publisher is commercial or funded.

Cardiff also has a compact market. Senedd coverage, city council services, university output and national broadcaster bureaux all interlink. A broken redirect on one site becomes a broken citation on another.

Bilingual content handling and hreflang for Cardiff sites

Welsh and English are both official in Wales, and search engines treat them as separate language versions. Google's own guidance on international and multilingual sites is the reference point for how to mark that relationship.

Use one URL pattern and keep it. A /cy/ prefix for Welsh and the bare path for English is the common Cardiff choice. Do not mix subdomains, parameters and prefixes across the same pair.

Machine translation is the trap. A Welsh page produced by an automated pass often reads oddly and duplicates the English text closely enough to compete with it. Commission real translation for service pages, policy pages and anything a citizen must act on.

Canonicals must not collapse the pair. Point the Welsh canonical at the Welsh URL, never at the English one. If you canonicalise Welsh to English you remove the Welsh page from the index and fail your own language duty.

Watch the language switcher. It should link to the equivalent page, not the homepage. A switcher that dumps users on /cy/ is a usability fault and a crawl signal that the pair is not really paired.

A bilingual content audit is the cheapest way to catch this. Export both language trees, match URLs by path, and list every page with a missing, one-way or self-referential hreflang value.

Welsh Government standards and public sector accessibility duties

Public bodies in Cardiff operate under a formal accessibility regime. The Public Sector Bodies (Websites and Mobile Applications) (No. 2) Accessibility Regulations 2018 set the duty for public sector websites and apps, and the Equality Act 2010 provides the wider duty not to discriminate against disabled people. GOV.UK's summary of accessibility requirements for public sector websites and apps sets out the duties that apply to UK public sector bodies.

The other standard is the language duty. The Welsh Language (Wales) Measure 2011 gives Welsh Ministers the power to set Welsh Language Standards, which the Welsh Language Commissioner oversees. For a public body those standards can require a Welsh-language version of a service page, so an English and Welsh URL pair is a legal obligation as well as a ranking decision.

The headline duty is a published accessibility statement, reviewed at least annually, naming known issues and how to report them. The Government Digital Service publishes the model statement and runs the monitoring programme that samples public sector sites each year, and the Equality and Human Rights Commission enforces the Regulations. That statement is also a content asset: it should be crawlable, linked from the footer, and available in Welsh.

The technical standard is WCAG 2.2 at level AA. The WCAG 2.2 specification is the version to cite in audits now, not 2.1. New success criteria in 2.2 cover focus appearance, dragging movements and consistent help.

Accessibility and SEO overlap more than teams expect. Descriptive link text, correct heading order, captions and transcripts, and a logical reading order all serve both. Alt text that describes the image serves both.

PDFs are the usual failure point. A scanned PDF with no text layer is invisible to screen readers and to search. If a document must stay as a PDF, tag it and give it a real title.

Third-party widgets are the second failure point. Cookie banners, chat tools and embedded maps often fail contrast or keyboard access, and the public body still owns the outcome.

Keep the audit trail. For each fix, record the criterion, the page, the person who approved it and the date. That record is what a reviewer asks for, and it is the same record a technical SEO planning template should hold.

Crawl, index and sitemap checks for large Cardiff content estates

Large archives fail quietly. A news archive with 200,000 URLs and a service directory with 40,000 pages both need crawl budget spent on pages that matter.

Run the audit with named tools and record which one produced each finding: Google Search Console for index coverage and query data, a crawler such as Screaming Frog or Sitebulb for URL-level and template-level issues, PageSpeed Insights or Lighthouse for lab checks on the main template types, and Ahrefs or Semrush for backlink and visibility context.

Start with the sitemap. Google's overview of sitemaps explains what a sitemap does and which files it accepts. A single sitemap file may hold up to 50,000 URLs and be no larger than 50MB uncompressed, so split by type, keep each file under those limits, and list only canonical, indexable URLs.

Then check for crawl traps. Faceted search, calendar pagination and print versions are the classic three on Cardiff media sites. Block the combinations, not the whole directory.

Audit internal search result pages. They generate infinite URLs, they are rarely useful in the index, and they compete with real content. A noindex, follow tag is usually right.

Check redirect chains. Site migrations, section renames and language moves leave chains of three or four hops. Each hop costs crawl and loses a little signal.

Check orphan pages. Any page with no internal link is found only through the sitemap, which means it is found slowly. Pull a crawl, compare against analytics, and link the winners.

Plan the work in a fixed order rather than reacting to whatever the last crawl shouted about. A repeatable technical SEO review methodology keeps the queue honest when two teams share one backlog.

Core web vitals and mobile-first checks for media publishers

Core Web Vitals are three field metrics: Largest Contentful Paint, Interaction to Next Paint and Cumulative Layout Shift. The Web Vitals definitions and thresholds give the pass, needs-improvement and fail bands for each. At the 75th percentile of page loads the pass bands are LCP within 2.5 seconds, INP within 200 milliseconds and CLS below 0.1.

Media sites feel all three. The LCP element is usually the lead image or the headline block. INP suffers when ad scripts and consent tools run on the main thread. CLS appears when ad slots resize after load.

Fix LCP first. Serve the lead image in a modern format, size it correctly, and preload it. Do not lazy-load the element above the fold.

Fix INP next. Defer non-critical third-party scripts, and load consent tooling before ad calls rather than after. Test on a mid-range Android device, because that is the field condition that counts.

Reserve space for every ad slot and embed. A fixed aspect ratio on the container prevents most layout shift. This matters more on Welsh pages if the translation runs longer and pushes content down.

Mobile-first indexing means the mobile render is the one assessed. Check that the mobile template carries the same hreflang tags, canonicals and structured data as desktop.

Public bodies have an easier time here. Fewer third parties, lighter pages. The risk is embedded video and mapping tools, which should be click-to-load rather than loaded on arrival.

Report field data, not lab scores. Field data from real Cardiff users is what a funder or an editor will accept, and it belongs on a shared technical SEO reporting dashboard alongside index coverage.

A Cardiff technical SEO checklist for media and public bodies

Work through this once a quarter. It assumes one English and one Welsh tree, a shared template, and at least one large archive.

Add these to the same quarterly pass:

A worked example

Take a Cardiff public body with a 12,000 page site, half in Welsh. A crawl finds 3,100 Welsh pages with no hreflang at all, because a template update dropped the tag.

The team adds reciprocal hreflang to the template, redeploys, and re-crawls. Welsh pages entering the index rise over the following weeks, and Welsh-language search impressions recover.

The same sprint fixes 240 untagged PDFs and reserves ad-free space above the fold, which lifts the mobile LCP into the pass band. One template change, three problems closed.

Sequence matters when several teams share a release train. A technical SEO tool implementation plan that names the owner, the environment and the rollback step is what stops a template fix from being reverted by the next deploy.

Common questions