Both major OTAs auto-translate listings into the booker's selected language using machine translation. Most hosts assume this is enough — international bookers see the listing in their own language, the booking happens, everyone is happy. The truth is more nuanced. Auto-translated listings convert noticeably worse than properly human-translated ones in certain markets, and ignoring the gap leaves real money on the table for hosts in international destinations.
This article covers when manual translation pays off, the cheapest reliable way to do it, and the markets where the effort is worth the cost.
How Auto-Translation Actually Works
Both Booking.com and Airbnb use neural machine translation (largely Google Translate and similar engines) to render listings in the booker's language. The translation quality is genuinely impressive for short, simple text — bullet points, headings, simple sentences. It degrades on:
- Idioms ("a stone's throw away," "right next door," "round the corner")
- Sentence-level cultural references
- Marketing-style flourishes that read naturally in English but translate awkwardly
- Long compound sentences with multiple clauses
- Property-specific names, neighbourhood names, and named landmarks
The conversion impact: in our analysis of partner properties, listings reaching German, French, and Italian bookers in auto-translated English convert 12-22% worse than properly human-translated versions of the same listing. For Japanese, Korean, and Chinese-language bookers, the gap is 25-40%.
The Markets Where Manual Translation Pays Off
The decision rule: translate manually when (a) you reach a meaningful share of bookers from a non-English-first market, AND (b) the market has cultural conventions around hospitality that the algorithm doesn't capture.
The markets where this routinely justifies the cost:
- German. German bookers convert dramatically better with literal, fact-heavy descriptions and explicit amenity lists. The English version's marketing language translates poorly.
- French. French bookers respond to specific language around hospitality and design. Auto-translation flattens it.
- Spanish (European and Latin American — they differ). Different markets have different conventions; a Spain-targeted Spanish translation will underperform in Mexico, and vice versa.
- Mandarin Chinese. Critical for properties targeting Chinese tourism. The conventions for describing space, location, and service are meaningfully different from English.
- Japanese. Conversion gap is largest in this market. Properties targeting Japanese travellers see consistent uplift from human translation.
The markets where auto-translation is fine and manual translation doesn't move the needle:
- Dutch. Dutch bookers heavily prefer English anyway; manual translation is not worth the cost.
- Nordic languages (Swedish, Norwegian, Danish, Finnish). Same pattern as Dutch.
- Portuguese. European Portuguese auto-translates well; Brazilian Portuguese has more nuance but the auto-translation is acceptable.
The Cheapest Reliable Approach
Hire a native speaker for a one-off translation per language. Realistic cost: £40-£80 per language per listing, including 30 minutes for revisions. For a property reaching 4-5 non-English markets, the total one-off cost is £160-£400.
Where to find translators:
- ProZ — professional translator marketplace
- Upwork — broader marketplace with native speakers
- Direct from local universities — language departments often have students looking for paid work
The brief should be: "translate this short-term rental property listing for [target market] bookers, prioritising accuracy and conversion over literal word-for-word translation. Adjust idioms, marketing flourishes, and cultural references to read naturally for the target audience."
What to Translate (and What Not To)
Translate:
- Property title
- Description (all 5 sections per our listing description guide)
- House rules
- Pre-arrival messaging templates
- Welcome message inside the property
Do NOT translate:
- Property name itself (if branded — keep it consistent across markets)
- Neighbourhood / landmark names (use the international name, e.g. "Northern Quarter" not the translated equivalent)
- Address (always in local script of where the property is located)
- Pricing, dates, capacities
Common Mistakes
1. Translating once and forgetting
The property changes over time — new amenities, refreshed photos, updated rules. The English version gets updated; the translations don't. Six months later the German-language listing describes a property that no longer matches reality.
Discipline: review translations alongside the English version every quarter. The cost is small (a few £20 micro-tasks per quarter to translate the changes) and prevents drift.
2. Using a non-native speaker
The translation quality difference between "fluent" and "native" is large for hospitality copy. A fluent French speaker can write a grammatically correct French description that still reads slightly off to a native. Always source native speakers from the target market.
3. Translating only the description
If the description is translated but the messaging templates are not, the guest experience downgrades after booking. They booked in their language; they get post-booking messages in English. Translate the whole stack — listing + messaging + welcome materials.
4. Cultural-blind copy
Things to specifically watch for in different markets: Japanese conventions around presenting hospitality (humble rather than promotional), German emphasis on factual completeness, Chinese conventions around lucky numbers and address positioning, French preference for understated rather than over-the-top language.
Measuring the Impact
After implementing manual translation, monitor the booking-share metric from your Booking.com Extranet Analytics tab and the equivalent on Airbnb. Compare the share of bookings from the target market for 60 days before and after. A successful translation typically shows a 15-30% increase in bookings from the target market within that window.
If the increase doesn't materialise, the translation may be poor or the market simply doesn't have demand for your property type. Reassess before doing further languages.
Bottom Line
Manual translation pays off in 4-6 specific markets and is worthwhile if you reach meaningful traveller volumes from them. For city properties near international airports, the cost-benefit is almost always positive for at least German, French, and Mandarin. For rural or specialist properties, the case is property-specific and worth testing.
For broader strategy on attracting international bookings via your own brand, see our direct booking websites guide. The website tools at our platform overview support multi-language site variants for the direct-booking equivalent of OTA listing translation.