The 8 languages that make the difference
Cross-referencing inbound tourism data for France (Atout France, INSEE 2024-2025) with what actually happens in dining rooms, here is the top 8:
- English — covers the UK, US and Australia, and serves as the Asian/European lingua franca (~35%)
- German — the largest European visitor group (~12%)
- Spanish — covers Spain and Latin America (~10%)
- Italian — strong cross-border traffic (~8%)
- Dutch — Belgium and the Netherlands, your closest neighbours (~7%)
- Chinese (simplified) — high average spend per table (~5%)
- Japanese — upmarket food-focused guests (~3%)
- Portuguese — Portugal plus Brazil (~3%)
Total: 80 to 85% of visitors to France read at least one of these 8 languages natively. The rest generally read English as a second language.
Why not more, why not fewer
Past 8 languages, the return on each additional one collapses. Russian, Korean and Arabic each account for 1 to 2% of arrivals. Very few restaurants can justify paying for quality translation there. Go below 6 or 7 languages, though, and you leave whole segments of visitors with nothing to read.
One caveat: a restaurant in Strasbourg should push German harder; one in Biarritz, Spanish; one in Nice, Italian. Tailor your top 8 to where you are.
The Google Translate errors that send guests elsewhere
Plenty of restaurants paste their menu into Google Translate and use whatever comes out. It's fast, but the result is often worse than a French-only menu. A few real examples:
- "Magret de canard" → "Duck magret" (instead of "Duck breast")
- "Tartare de bœuf" → "Beef tartar" (instead of "Beef tartare" — different spelling entirely)
- "Carpaccio de Saint-Jacques" → "Saint-Jacques carpaccio" (instead of "Scallop carpaccio")
- "Aile de raie aux câpres" → "Wing of stingray with capers" (technically accurate, but off-putting)
- "Lapin à la moutarde" → "Rabbit in mustard" (instead of "Rabbit with mustard sauce")
The outcome: your English-speaking guest misreads the dish, or hesitates and orders the Caesar salad to be safe. You sell less, and your menu looks amateurish.
"We ran our menu through Google. A Japanese guest very politely pointed out that we had written 'sauce of death' instead of 'house sauce'. Mortifying." — Pierre, chef in Aix-en-Provence
Food-trained AI vs raw machine translation
The gap between a generic translation engine and an AI trained on food comes down to three things:
- Culinary context: a food-trained AI knows that "magret" becomes "duck breast" in English, "Entenbrust" in German, "petto d'anatra" in Italian — not "magret" everywhere.
- Names that stay put: "tartiflette", "andouillette" and "ratatouille" aren't translated. You keep the name and add a short description.
- Borrowed terms: "carpaccio" is Italian and stays Italian, whatever the target language.
A tool like QRMENUPRO translates each dish with that context built in. You read it over and approve — about 30 seconds per language — and it's live. More on our features page.
Beyond translation: the multilingual experience
Translating the menu isn't the whole job. A proper multilingual experience also means:
- Automatic browser language detection (your American guest lands on the English menu without clicking anything)
- A clear language switcher at the top of the menu
- Allergens translated too (a German guest is looking for "Erdnüsse", not "cacahuètes")
- Prices in euros, with a note if you accept international payment methods
- A short description of the restaurant in each language ("Family restaurant since 1987")
Key takeaways
- 8 languages = 80-85% of visitors covered
- The list: English, German, Spanish, Italian, Dutch, Chinese, Japanese, Portuguese
- Adjust for your location (border town, seaside, city centre)
- Raw Google Translate = a steady source of embarrassing errors
- Food-trained AI = culinary context respected
- Auto-detection + clear switcher + translated allergens = the full experience