Why "Just Translate the Website" Fails as a China Strategy

Someone quoted you for a Chinese version of your website. You paid, it shipped, and the translation is good — a native speaker checked it. Six months on, those pages have a few hundred visits, most of them bots, and not one enquiry you can trace to them. The natural conclusion is that the translation must have been bad, or that Chinese travelers simply aren't interested in what you sell. Usually neither is true. The translation did exactly what translation does. The four things standing between a Chinese traveler and your booking were never language.

The short version: translating your website solves the language barrier and leaves four others untouched — discovery (Chinese travelers don't start at your domain), delivery (the page loads badly or not at all from the mainland), relevance (it answers a Western buyer's questions, not a Chinese one's), and the reply path (a form landing in an inbox nobody watches). A translated site is a useful artifact wired to the wrong plumbing. Fix the plumbing on two or three products before translating the other forty.

Nobody was going to arrive at your domain anyway

The Western funnel assumes a search engine sends someone to your website, where your website does the selling. Chinese discovery mostly doesn't leave the app it started in. People search inside Xiaohongshu and Douyin, ask inside a WeChat group, and open a Mini Program rather than a browser — the mechanics we covered in how Mini Program discovery actually works and why good Xiaohongshu notes keep selling for months. Recommendations get passed around in group chats you'll never see. Your domain is not a stop on that route.

So be blunt about the causality, because it's usually sold backwards: a Chinese page does not create visitors. If nobody from China is reaching you today, translating the site changes nothing about that. It changes what happens to the ones who already are — which is worth real money, but it is a different claim, and you should insist on hearing which one you're buying.

The page has to actually load

This is the failure nobody checks, because it's invisible from your office. Web fonts served from a CDN that resolves slowly, an embedded video from a platform that won't play, a map widget that hangs for thirty seconds, third-party analytics blocking first paint, hero images on a host that never answers. A page that renders in a second at home can take fifteen from a phone on mainland mobile data, or never finish at all.

Half-loading is worse than slow. The traveler sees a broken layout with missing images and draws a conclusion about how carefully your company does everything else.

We hit this concretely on a tour catalogue we rebuilt recently: every product image had to be re-hosted on a domain we control before the pages were usable in the mainland, because the originals sat on a booking platform's asset host that simply didn't respond. Nothing about that problem was visible from a browser in Europe. Test from where your customers are, not from where you are.

It answers the wrong questions

Translation faithfully preserves the questions your English page chose to answer — and those were chosen for a buyer with different assumptions and different anxieties. The gap isn't vocabulary, it's editorial, and it connects to the broader pattern in what replaces reviews and star ratings.

• What exactly is included. Inclusion lists get read closely, line by line. "Full-day experience" is a category, not a product description.
• Whether anyone speaks Mandarin. If nobody does, say so plainly. Silence reads as yes, and produces a complaint on the day rather than a lost booking beforehand.
• What the meeting point looks like. An address is not enough. A photo and a nameable landmark are, because that screenshot gets shown to a taxi driver who doesn't read your language either.
• What happens if it rains, if they're late, if they cancel. The cancellation and deposit terms are a trust signal, not fine print — see refunds, deposits and no-shows.
• Whether it works for three generations travelling together. Party composition skews differently, and "suitable for families" doesn't answer it.

A translator adds none of these, correctly. They weren't on the page to translate.

The reply path is where it usually dies

Say the page loads and reads well. The last step is still a contact form that sends you an email — and Chinese customers largely don't run their lives out of an inbox. Some will fill the form anyway, and then never see your reply, because the address they typed belongs to an account opened years ago to register for something.

Two things make the path work. The enquiry should arrive structured — intended date, number of people, which product, and a contact route the person actually uses — so you can quote without three rounds of clarification. And it should reach someone who answers the same day, because the enquiry was sent from a phone in a moment of interest that will not last a week.

Then plan for the part that proposals leave out. A traveler will type a paragraph of free-form Chinese into the message box, and it will land in your inbox in Chinese. Structured fields cover most of an enquiry, never all of it. Decide in advance who reads that paragraph and how — machine translation, a bilingual contact, or fields designed to capture the question so it doesn't need to be typed as prose. It's a small problem if you've planned for it and an embarrassing one if you haven't.

What a translated page is genuinely good for

None of this makes the Chinese page worthless. It makes it a destination rather than a strategy, and destinations matter when someone is already on the way:

• Somebody already has your link — from a group chat, a guide, a hotel desk, or a QR code on your own counter.
• A partner agency needs something to show a Chinese client, which is still how a good share of demand moves — see how Chinese group tours actually get booked.
• You need one canonical place where the facts are right, so the same numbers appear in every chat and survive being screenshotted.

Translate because people will arrive. Don't translate expecting arrival.

A cheaper order of operations

1. Pick two or three products, not the catalogue. Your best sellers, where you already know the questions people ask.
2. Write them, don't translate them. Answer the five questions above, in that order of prominence.
3. Host them so they load from a Chinese network — images and fonts included, third-party embeds removed rather than optimised.
4. End with a structured enquiry that routes to a channel a human reads within hours.
5. Tag the source so you can see which product is being read and where the reader came from. Note the honest limit: that tells you which page and which product, not the reader's nationality.
6. Then, and only then, think about discovery.

Steps 1 to 5 change what happens to demand you already have. Only step 6 changes how much demand exists — and that's a slower, more expensive project, which is exactly why it should be funded by results from the first five rather than promised alongside them.

There's a quieter benefit to this order. If two or three properly-built products produce nothing across a season, you've learned something cheaply and truthfully: no Chinese demand is reaching you yet, and no volume of translation was going to invent it. Much better to learn that on three products than on forty.

Where CN1X fits

We build the Chinese pages and Mini Programs — written as answers rather than translations, hosted so they load in the mainland with the images included, ending in a structured enquiry that reaches your inbox with the date, the party size and the product already filled in, tagged so you can see which product drew it.

Enquiries go to your team. You keep the customer, the pricing and the payment, and you answer in the language you already work in. We don't promise you traffic — we make the arrivals you already have countable and answerable. See what that involves, or send us two products and we'll tell you what they'd look like in Chinese.