Website strategy · multilingual search
If your website attracts enquiries in more than one language, translate a complete route to becoming a customer before translating your entire archive. A useful service page, relevant proof and a working enquiry path are a stronger first investment than dozens of disconnected translations.
Choose a customer task, not a flag
A company serving Brussels may need Dutch, French and English content without selling three different services in three different countries. Start with the combination that matters: one customer need, one service and one language you can genuinely support.
Language is not a reliable proxy for location. W3C notes that a visitor’s location can differ from their preferred reading language and recommends easy-to-find language choices. Use labels such as Nederlands, Français and English rather than treating national flags as languages. W3C’s guidance on translated pages explains that distinction.
Before commissioning copy, collect evidence: the language of recent qualified enquiries, recurring questions on calls, relevant search queries where available, and the words customers actually use for the service. Country-level traffic alone cannot tell you which language those people prefer. Low traffic to an untranslated service is also weak evidence of low demand: you may not yet have a suitable page.
Validate the search task, not just the dictionary equivalent. Shopify’s multilingual SEO guidance distinguishes language targeting from country targeting and recommends checking local search terminology with native speakers. For a service business, that means checking whether customers want a quote, an emergency response, a comparison or an ongoing contract before writing the landing page.
Spend a small budget on a complete route
Imagine a Brussels maintenance company with a Dutch website, repeated French enquiries for office cleaning, and occasional English questions about a service it no longer offers. The company can afford twelve hours of translation, review and implementation. Those are planning assumptions—not an industry benchmark.
- 2 hours: confirm the French office-cleaning offer
- Check customer terminology, service area, scope and who will answer enquiries. Resolve unclear claims in the original before translating them.
- 4 hours: write the service and proof content
- Explain the work, exclusions, quotation process and a relevant project example. Proof may sit on the service page; it does not automatically need another URL.
- 3 hours: complete the enquiry route
- Include navigation, form labels, validation messages, privacy information, confirmation and the first reply. Coordinate any professional review the information requires.
- 3 hours: review and test the whole journey
- Have a fluent reviewer check meaning, follow the route on a phone, verify language links and assign an owner for future updates.
This budget buys one coherent French service journey. It does not buy a translated blog archive, a new country site or an English page for an unavailable service. If the actual work takes longer, reduce the launch scope or increase the budget; do not remove the review to meet the example.
Two decisions to settle before expanding
Should I translate my entire website at once?
Not necessarily. Start with one commercially relevant service and a complete route to enquiry in the chosen language. Expand when demand and the ability to maintain the content justify it.
Does a French page need to target France?
No. Language and country are separate decisions. A French-language page can serve French-speaking customers in Belgium or elsewhere; use country-specific targeting only when the offer genuinely differs by market.
Follow the language past the send button
The service page may read naturally while its next step breaks the promise: a French visitor reaches a Dutch form, encounters an English error and receives an unexplained confirmation. Count that as an incomplete launch, even if every headline is translated.
Walk through a specific enquiry with a fluent reviewer. Can the visitor establish whether the service fits, understand what information is needed, correct a mistake and know what happens next? State the languages available for the actual conversation. Do not imply full French support if only the website copy is French.
Some destinations will remain untranslated during a staged rollout. Make that boundary visible before the click—for example, “Technical specification (English)”—rather than silently moving the visitor into another language. A language switch on a service page should preserve the topic where an equivalent exists. Where it does not, offer an honest route to available content; do not label an unrelated homepage as the translation.
This is a content-architecture decision before it is a translation task. A search-led website architecture can map the relevant service, proof and enquiry destinations before copy is commissioned.
Give implementation a clear handover
Ask for separate, crawlable URLs for the language versions; visible content and navigation in the intended language; working links between equivalent pages; and appropriate language annotations. Avoid a setup that shows different languages only after a visitor changes a browser setting, or forcibly redirects someone away from the language they chose. These points follow Google’s multilingual-site guidance.
For each proposed URL, record the customer task, language, corresponding original, next-step destination and person responsible for updates. That simple handover exposes missing destinations before they become public dead ends. Technical annotations cannot repair an absent translation or an unclear offer.
Expand when the evidence supports another journey
After launch, separate discovery from business value. Review relevant search queries and landing-page visits where data is available, then qualified enquiries, the service requested and whether the team could serve the customer in that language. Use the language explicitly chosen by the customer when possible; a browser setting is not a reliable substitute.
Do not promise a fixed traffic uplift. Ahrefs’ December 2024 multilingual analysis uses a small sample of major global brands. It illustrates the scale multilingual content can reach; it does not isolate translation as the cause or forecast results for a Belgian service company.
Choose a review period that covers your buying cycle and enough real enquiries to learn from. If visits arrive but enquiries fail, inspect the offer and handover before translating more pages. If relevant enquiries grow and the content remains maintainable, consider the next service-language combination. If evidence is sparse, record uncertainty rather than calling the experiment a success or failure.
The next useful action is to map one journey. Pick the service, identify the customer’s language, connect the proof and enquiry steps, and name the person who will keep the facts aligned. If the customer vocabulary is unclear, semantic analysis for service-page intent helps define what the translated page must answer.