Translation is a market investment, not a website decoration
A translated website changes words. A localised website changes the buying experience so it works in a specific language or region. That may include terminology, search demand, proof, currency, measurements, payment or enquiry flow, delivery promises, policies and the way a customer reaches sales or support.
The commercial question is therefore not “Can we translate the site?” Modern tools make that easy. The question is whether a maintained local experience will help suitable buyers discover, understand, trust and buy from a business that can actually serve them.
My verdict is: localise the smallest complete journey that can test a real market hypothesis. A market-specific landing page and decision path often create more useful evidence than an automatic translation layer across hundreds of pages. If the market, offer or operating model is still uncertain, use the new-market evidence guide first.
This intent is also distinct from deciding whether the market needs a separate brand. Translation may support one shared brand, an endorsed market identity or a separate business, but it cannot make that architecture decision by itself.
The Market-Language Website Readiness Gate
I would require six connected answers before funding a multilingual or multi-regional rollout.
Is the market real?
Use buyer research, enquiries, partners, sales conversations, search behaviour and small tests—not population size.
What blocks the decision?
Identify whether buyers need another language, local terminology, regional detail or all three.
Does the offer travel?
Confirm pricing, margin, positioning, proof and customer value for the specific market.
Can the promise be kept?
Test sales coverage, payment, fulfilment, support, returns, lead routing and regulatory requirements.
Will buyers trust it?
Adapt evidence, policies, contact routes and decision information without pretending to be local.
Who keeps it accurate?
Assign source content, translation review, approvals, publishing, measurement and future updates.
The US International Trade Administration separates website work into internationalisation, regionalisation and localisation. Its model is commercially useful: first make the site structurally ready, then test a regional experience, and only then accept the higher cost of country-specific localisation. A business does not need to jump from one English website to a fully translated global estate.
Translate meaning, then localise the decision path
Direct translation is necessary when the customer cannot understand the source language, but it is not sufficient. The UK Government's export guidance says adaptation also involves cultural expectations, locally appropriate content, currency, payment, delivery, returns and search optimisation. The same principle applies to B2B: procurement language, certifications, service territory, implementation, response times and proof may matter more than literal word choice.
| Layer | Owner question | Useful evidence |
|---|---|---|
| Language | Can the buyer understand the offer accurately? | Search terms, sales language, customer interviews, specialist review |
| Market | Does this page reflect how buyers choose locally? | Competitor journeys, objections, proof expectations, procurement stages |
| Operation | Can the business fulfil what the page promises? | Coverage, stock or capacity, payment, delivery, support and policy readiness |
| Discovery | Can search engines find the correct version? | Stable URLs, crawlability, internal links, sitemaps and locale annotations |
| Commercial value | Is the local journey producing worthwhile customers? | Qualified demand, sales progression, contribution and maintenance cost |
Google distinguishes multilingual sites from multi-regional sites. It recommends separate URLs for language versions and says visible page content—not a URL or code attribute alone—is used to determine language. Google also supports hreflang annotations so corresponding locale pages can be understood as alternatives. Those are implementation safeguards after the market decision, not reasons to translate.
Choose the smallest defensible rollout
| Evidence pattern | Decision | Next move | Avoid |
|---|---|---|---|
| Demand unclear; no local operating path | Research | Interview buyers and test one market proposition | Translating the whole website for reach |
| Same language; different buying context | Regionalise | Create one market page with local proof, offer and route | Duplicating pages with only spelling changes |
| Language barrier; demand and delivery proven | Localise | Build a complete high-value journey with qualified review | Publishing raw machine translation |
| Several markets; limited maintenance capacity | Sequence | Prioritise by customer value, evidence and operating readiness | Launching every language simultaneously |
| Translated traffic; weak enquiries or orders | Diagnose | Inspect fit, offer, trust, route, sales and fulfilment | Adding more translated content automatically |
A manufacturer entering a new geography may begin with one local buyer page, sector proof, enquiry route and sales owner rather than translating its entire technical library. The manufacturing market-entry guide shows how offer fit, evidence and sales progression belong together. An ecommerce business may need product, price, payment, delivery and returns to work as one journey before local traffic is valuable.
The Search-to-Localised-Revenue Evidence Loop
A successful rollout connects discoverability to customer value while keeping every market version accurate.
Choose one market
Define the buyer, problem, offer, route to market and evidence that justifies a test.
Map one complete journey
Select the pages, proof, policies and conversion path needed for a real decision.
Adapt commercial meaning
Use tools for speed, then apply in-market, legal and specialist review where needed.
Make versions discoverable
Use stable URLs, obvious language, navigation, canonical logic and the appropriate locale mapping.
Connect demand to sales
Route enquiries or orders with market, language, offer and source context intact.
Judge customer value
Compare qualified demand, contribution, delivery friction and maintenance cost by market.
Review AI Growth and discovery, landing-page work, conversion tracking, case evidence, evidence standards, the Google and AI discoverability guide and Thomas's operating model.
Practitioner note: I would not ask a translator or AI tool to decide which market deserves investment. The business must own the market hypothesis, offer, proof, delivery and success definition. Translation then helps execute a decision whose commercial logic already exists.
Sources and evidence notes
Sources and current search results were checked on 14 October 2026. Search prioritisation is qualitative; no unverified search volume, conversion benchmark, traffic threshold, forecast or client result is used. The Market-Language Website Readiness Gate, Search-to-Localised-Revenue Evidence Loop and decision matrix are original ThomPerformance practitioner analysis.
Frequently asked questions
Should a business translate its entire website at once?
Usually not. Start with the pages that complete one valuable customer journey: market-specific landing page, offer or product information, proof, pricing or enquiry path, and the policies customers need to decide. Expand only when search, sales and customer evidence justify the maintenance cost.
Does a translated website need a new country domain?
Not automatically. A subdirectory, subdomain or country-code domain can work, depending on brand architecture, local eligibility, operating commitment and technical resources. The important decision is whether each locale has a stable, crawlable URL and can be maintained correctly—not whether it looks locally impressive.
Can AI translate the website without human review?
AI can accelerate a first draft and terminology workflow, but it should not own the final commercial meaning. A qualified reviewer should check buyer language, claims, terminology, legal or regulated content, cultural fit and the next action. The risk is not awkward prose alone; it is a wrong promise or unusable journey.
Is localisation needed when the new market uses the same language?
It can be. Buyers may use different search terms, spelling, currency, measurements, evidence, service expectations, pricing conventions and regulatory language. A US, UK and Australian page can share a language while still needing different commercial details.
How should the business measure whether localisation worked?
Measure the whole market journey: suitable search visibility, engaged visits, enquiries or orders, qualification, sales progression, contribution, service or return issues, and the ongoing cost of keeping every version accurate. Traffic growth alone does not prove a viable new market.
Market-discovery diagnostic
Build the local journey only when the market can support it
Share the target market, offer, current international demand, website structure and delivery model. I'll identify whether to research, regionalise one journey, localise a proven path or postpone the investment until the commercial foundation is stronger.
Discuss a website localisation decision