Picture a founder running Meta ads into Germany. The dashboard fills with clicks; the inbox stays quiet. So she opens her own site the way a German buyer would, clicks the little flag, lets the translation widget do its work, and finds her pricing page rendered in German that reads like a robot filled it in under duress. She didn't enter the German market. She announced she wasn't really in it.

That is the quiet trap of the multilingual website. Owner-operated European businesses get told that going European means switching on six languages. It doesn't. More languages rarely mean more markets — they mean more half-built versions of one site, none of them good enough to win a buyer. The real question was never how many languages you offer. It's which markets earn a proper one.

TL;DR: A multilingual website helps only when each language is fully localised and kept current. For most owner-operated European businesses that means picking one or two markets worth a real, hreflang-correct experience, and serving everyone else in clean English on purpose, rather than bolting six machine-translated versions onto one site.

Should your website be multilingual at all?

Only if you can localise and maintain each language properly. For most owner-operated European businesses, the honest answer is one or two languages, not six.

The case for more than English is real. Across the EU, only about 47% of people speak English as a foreign or second language, per the European Commission's 2024 Eurobarometer on languages. So an English-only site does leave a large share of the continent reading something they're not fully comfortable in. But that fact is not an argument for translating everything. The EU has 24 official languages. You are not going to serve all of them, and you shouldn't pretend to.

The decision is market-by-market, not language-by-language. Each language you add is reach gained and upkeep owed — forever. Treat "multilingual" as a setting you flip on and you've already made the mistake. Treat it as a commitment you take on for specific markets, and you'll make a far smaller, far better list.

Why a translation plugin makes you look smaller, not bigger

A one-click translation widget doesn't open a market — it tells a native buyer you couldn't be bothered to show up properly in theirs. Language matters enormously to buyers, which is exactly why machine translation backfires.

The numbers are blunt. CSA Research's "Can't Read, Won't Buy" survey of 8,709 consumers across 29 countries found 76% prefer to buy when the information is in their own language, and 40% will never buy from a website in another language. Those same buyers can smell auto-translation in a sentence. A German shopper who hits a checkout where half the field labels are still English doesn't think "close enough" — they leave.

Google agrees, in its own way. Its guidance treats unreviewed machine-translated content as low quality, and that can drag down rankings across your language versions, not just the bad one. Localised pages are only seen as duplicates when the main content stays untranslated, so a widget that swaps a few visible strings and leaves the rest gives you the worst of both: it looks translated to a crawler and reads broken to a human. You wanted to look bigger. You looked careless in two languages instead of competent in one.

Which European markets actually deserve a localised site

Start from where the money and the demand already are, not from a map of Europe. The markets that earn a real localised site are the ones already pulling revenue, enquiries, or ad spend — not the ones that look good on an expansion slide.

Run three questions over each candidate market:

• Is real revenue or qualified demand already coming from here?

• Are you spending ad budget to reach it?

• Can you keep every page, and every new post, current in that language indefinitely?

Two clear yeses, and the market earns a full localised experience. Anything less, and it gets clean, confident English. That is not a cop-out. Seven in ten young Europeans can hold a conversation in English, per the same Eurobarometer data, and in B2B and SaaS the buyer reading your site is very often one of them. A B2B firm whose pipeline is 70% German-speaking should build a real German site; the thin trickle from Italy does not justify a second one you'll never maintain. Pick the market that pays. Serve the rest in a language you can actually keep sharp.

How to localise the markets you pick without breaking your SEO

Localisation is not translation. Translation swaps words; localisation rebuilds the page so it sells to that specific buyer. Get that distinction wrong and even a "properly translated" site underperforms.

Real localisation adapts the copy, the examples, the proof, and the commercial framing: VAT and pricing shown the way that market expects, not a converted afterthought. It also respects the mechanics: German runs roughly a third longer than English and will break a layout that was only ever tested in short English strings. Then comes the engineering most plugins skip. Give each language its own URL, add hreflang annotations so Google serves the right version to the right reader, and don't auto-swap content based on browser or cookie alone — those are Google's own recommendations for multi-region sites, and getting them wrong quietly buries the version you spent the most on.

This is the unglamorous part where a senior pair of hands earns its place. In our experience, the damage in a multilingual build is almost never the words — it's a hreflang error sending German searchers to the English page and halving the conversion you paid a translator for. A conversion-first website treats those signals as part of the build, not a plugin you bolt on after. If you want to see who's actually doing that work, the About page is a fair place to start.

Every language you add is a weekly commitment, not a one-off

"We'll add the languages once and leave them" is exactly how you end up with six stale websites. A localised site is not a translation project that finishes — it's a second, third, and fourth site that now needs everything the first one needs.

Every new blog post ships times N. Every price change, every campaign landing page, every fixed typo, every updated phone number — times N. A language you can't keep current decays into a worse signal than no language at all: a French page still advertising a 2023 offer tells a French buyer you left. Consider the Berlin B2B owner who paid an agency to "add French and Italian," and now files a support ticket every time he wants to change one line, because he can't touch three frozen sites himself. He didn't buy two markets. He bought two liabilities.

This is the cost translation vendors never put on the invoice, and it's why a content engine that ships in the languages you chose beats a pile of languages nobody updates. Fewer markets, kept current every week, will always outsell more markets left to rot.

A 10-minute test before you add a language

Before you switch on another language, answer these in order. Stop at the first no.

• Where do your revenue and qualified enquiries actually come from today?

• Which markets are you already paying ads to reach?

• Will the copy be localised by someone who knows the market, not a plugin?

• Can you commit to updating every page and every new post in that language every week, indefinitely?

• Do you have separate URLs and correct hreflang in place, or a partner who'll set them up?

Can't reach a yes on upkeep? That market gets clean English, and you keep one site you can actually ship. The point isn't to think small. It's to be genuinely present in the markets you claim, instead of vaguely absent in six.

Going multilingual is an operating decision, not a plugin. Pick the markets that earn a real, localised, hreflang-correct site, ship them every week, and serve the rest of Europe in English you're proud of. That's the version that compounds.

Plan. Build. Iterate.

Want an honest read on which markets your site should actually serve? Book a 30-min call — every price is on the Plans page, no annual contract.