Technical SEO Consultant, Belgium

Language switchers, slugs and mixed-language templates

Multilingual problems in Belgium usually start in the CMS, not in the content. A common setup runs French and Dutch from one installation, with English added later and German for a handful of pages. Each CMS treats missing translations differently: some fall back silently to another language, some show an empty page, some publish French text under a Dutch URL. Search engines then see pages whose language does not match their address or their hreflang.

What I check on each template:

• Fallback behavior: whether an untranslated page returns content in the wrong language, an error or a redirect, and whether that matches what you intend.
• Language switchers: whether they point to the equivalent page or drop visitors on the homepage of the other language.
• Mixed templates: navigation, breadcrumbs, filters, cookie text and footers that stay in one language on another language's pages.
• Slugs: whether URLs are translated, and how accented characters such as é or ë are handled. Encoded characters work, but plain-letter slugs written consistently are easier to share and less fragile.
• Lang attributes: the html lang value on each page, which should match the content rather than the site default.

Each finding becomes a ticket naming the template, the CMS setting or code path responsible and a test the developer can run to confirm the fix.

Hreflang at scale: sitemaps, canonicals and partial translations

Hreflang across four language versions creates a large number of annotations, and they fail quietly. When the CMS generates them automatically, typical faults are missing self-references, return links pointing to redirected URLs, annotations listing a translation that was never published and canonicals that send every version to the French page. Search engines then ignore part of the set and may show French pages to Flemish searchers or the reverse.

For larger Belgian sites, a cleaner approach is to move annotations out of page heads into XML sitemaps generated from a single source of truth: a table recording, for each piece of content, which language versions exist and at which URLs. That table then drives the sitemap, the switcher and internal links, so they cannot disagree. The checks that follow:

• Every annotation returns, uses valid codes such as fr-BE, nl-BE, de-BE or en, and points to an indexable, canonical URL.
• Pages that exist in only some languages list only those partners.
• An x-default is chosen deliberately, often a language selection page or the English version.
• Sitemaps are split by language so Search Console coverage can be read per version.

The strategic side, such as whether Belgian French should be kept apart from French for France, belongs to my international SEO page for Belgium.

Brussels document libraries, PDFs and crawl attention

Brussels-based associations, federations, consultancies and public affairs firms publish a steady stream of documents: position papers, consultation responses, reports, newsletters and event material, frequently in several languages. These files can be valuable in search, because people look for them by title or topic, yet they create technical problems when nobody manages them.

• The same PDF uploaded several times under different URLs, or linked from a news item and a library page with different parameters.
• Outdated versions still indexed and outranking the current one.
• PDFs with no title metadata, so search results show a file name.
• Document pages reachable only through an internal search box, which crawlers do not use.
• Members-only content exposed by accident, or public content hidden behind a login.

I map the document library and decide which files should be indexed, which should get an HTML summary page that links to them and which should carry a noindex header or be removed. Server logs, where available, show how much crawler activity goes to old files compared with current pages. For organizations whose website also handles memberships, events or grant work, my digital transformation page for Belgium looks at the systems behind it.

Structured data for bilingual addresses and local businesses

Brussels is officially bilingual, so most street and municipality names there have a French and a Dutch form, and businesses elsewhere in Belgium may publish addresses in more than one language too. That is perfectly normal for people, but structured data, business profiles and directory listings should be consistent within each language version so search engines can tie them to one entity.

My approach to Belgian markup:

• Organization markup with the legal name, trading names, enterprise number and VAT number where you choose to publish them, identical across language versions.
• LocalBusiness markup on each language version using that language's form of the street and town name, with the same postal code and coordinates.
• Opening hours, contact points and available service languages expressed in markup, so a Dutch-speaking visitor can see that Dutch support exists.
• Product and offer markup with euro prices consistent with the visible page and the feed for each language.

I check the markup against the visible page, since disagreement between the two is a frequent reason rich results are not shown. Markup supports eligibility for search features; it never guarantees them. The full method is on my schema markup consulting page, and the SEO consultant page for Belgium covers local listings and reviews.

Consent banners in several languages, tag weight and handover

A Belgian site may need its consent banner in French, Dutch, German and English, and some consent platforms load a separate script or text bundle for each. Add a tag manager, analytics, advertising pixels from campaigns in each language and a chat widget, and the scripts can outweigh the page itself. Core Web Vitals field data by template and language version shows whether that weight reaches real visitors, and which pages suffer most.

The review covers how the banner loads and whether it shifts the layout, which tags fire on every page although they serve one campaign, font files loaded for accented characters and images on templates shared across languages. The wording of the banner and the decision about which tags may run before consent are for your privacy advisor, with guidance from the Belgian data protection authority; my work is the performance side.

Implementation stays with your developers or agency, whichever language they work in. I deliver findings in English as tickets with evidence, the proposed change and a test, and I review fixes on staging and production. Migrations, such as merging separate French and Dutch domains into one site, follow my SEO migration process with a redirect map per language. All work is remote, with calls in Belgian business hours. The Belgium overview lists related services.