English, BM or Chinese: Choosing Content Language

Almost every Malaysian business asks this question the same way: “Should we do the website in three languages?” It sounds like a translation question. It is not. It is a budget question wearing a translation costume, and the honest answer is that most companies cannot afford to feed three content tracks properly in year one.

What follows is a way to decide. Not a list of translation tips, but criteria you can apply to your own numbers, so you leave with one language chosen and a clear signal for when to add the second. It sits under our content marketing work, where the content language Malaysia question comes up in nearly every first session we run at IZILI Digital Marketing.

Before the data, here is a practical walkthrough of how search engines handle sites that publish in more than one language.

PART 1 · DIAGNOSE

Which Language Do Your Buyers Already Search In?

IN BRIEFStart from evidence you already own. Read the actual words in your WhatsApp enquiries, Google Business Profile questions and Search Console queries. The content language Malaysia buyers use when they are ready to pay is usually narrower than the language mix of the country as a whole.

Your own inbox is the cheapest research you will ever run. Open the last fifty enquiries and sort them by the language the customer opened in. Then do the same for the ones that became paying jobs. Those two piles are often different, and the second pile decides your content language.

Four sources worth reading before you decide:

• WhatsApp and enquiry form text. Not what your staff replied in, but what the customer typed first, unprompted.
• Search Console query report. Filter for BM terms such as harga, servis and berdekatan, then compare impressions against your English queries.
• Google Business Profile questions and reviews. Reviews are unfiltered language evidence, and they often skew more local than your website traffic.
• Sales call notes. Note the language a buyer switches into when they start negotiating price. That switch is a strong signal.

One warning. Traffic that arrives in English does not prove English is the right choice. It may only prove English is the sole language you have ever published in, and absence of BM traffic on a site with no BM pages is not evidence of absent BM demand.

BENCHMARK BRIEFING 1 OF 4

What Does Malaysia’s Audience Look Like by Language?

IN BRIEFMalay speakers make up the largest share of citizens, English proficiency is unusually high by Asian standards, and internet access is near universal. That combination is why a single-language site can still perform here — and why the content plan behind it matters more than the translation.

Three official datasets set the boundaries of this decision. The Department of Statistics Malaysia puts Malays at 58.3% of citizens in Q1 2026, Chinese at 22.1%, Other Bumiputera at 12.3% and Indians at 6.5%. Household internet access reached 96.8% in the DOSM ICT Use and Access Survey 2024. And Malaysia scored 581 in the EF English Proficiency Index 2025, its highest band yet, with Kuala Lumpur at 588.

Read the last two columns as planning assumptions, not measured language preference. Malaysia does not publish a national survey of what language people prefer to read marketing in, and ethnicity is only a rough proxy for it. A Chinese-Malaysian engineer in Bangsar and a Chinese-Malaysian trader in Klang may sit in the same row and read completely differently.

PART 2 · DIAGNOSE

Why “Everyone Here Speaks English” Is a Costly Shortcut

IN BRIEFUnderstanding a language and buying in it are different behaviours. High English proficiency means Malaysians can read your English page. It does not mean English is the language they trust enough to send money in — and that gap is where quiet revenue leaks out.

The shortcut is understandable. Malaysia ranks in the high-proficiency band, business runs in English, and your own network replies to you in English. So the English-only site feels safe.

The problem is that comprehension and conversion are not the same test. A reader can follow your English service page perfectly and still hesitate at the payment step. The words that carry reassurance, such as warranty terms and refund conditions, land differently in a second language. That hesitation rarely arrives as a complaint. It arrives as a page with traffic and no enquiries.

This is also where budget quietly evaporates. Businesses answer weak results by publishing more English pages, when the real fix was one well-made page in the other language. It is the same trade-off we work through in content refresh versus new posts: more output is not automatically more return.

BENCHMARK BRIEFING 2 OF 4

How Much Does Language Change Buying Behaviour?

IN BRIEFThe largest published study on this found language preference changes purchase decisions, not just comprehension. Roughly three in four shoppers prefer buying where product information is in their own language, and four in ten will not buy from another-language site at all.

The figures below come from CSA Research’s Can’t Read, Won’t Buy study, run with Kantar across 8,709 verified consumers in 29 countries. Malaysia was not one of the 29 markets sampled, so treat these as directional evidence about how language behaves, not as Malaysian conversion rates.

The 65% line is the one worth sitting with. It says a plain, slightly awkward page in the reader’s language often beats a polished page in their second language — which is the opposite of how most Malaysian businesses budget for translation.

PART 3 · DESIGN

How to Pick Your First Content Language

IN BRIEFChoose on three criteria: who your highest-value buyer is, how contested that language’s search results already are, and which language your team can sustain without outsourcing the trust layer. Whichever wins on all three goes first, and everything else waits.

The trap here is picking the language with the biggest audience. A content language Malaysia decision made on population share alone almost always over-serves reach and under-serves revenue. A specialist B2B supplier with eight clients a year does not need the largest audience in Malaysia. It needs the right forty people.

BENCHMARK BRIEFING 3 OF 4

What Does Each Language Track Cost to Run?

IN BRIEFA language track is not a one-off translation bill. It is a recurring commitment of writing, editing, review and reply capacity. The model below shows why running three tracks at once costs far more than three times the attention of running one.

The figures are an illustrative planning model, not quoted prices. They exist so you can sanity-check whether your current budget can genuinely feed the number of languages you are considering.

Notice the column most plans forget: reply capacity. Publishing in Bahasa Malaysia creates BM enquiries, and a BM enquiry answered three days later in English undoes the work of the page that produced it.

PART 4 · DESIGN

Translate the Page, or Write It Fresh?

IN BRIEFTranslate the pages where the facts do the selling, and write fresh where the words do. Specification pages, policies and product details translate cleanly. Service pages and blogs usually need rewriting, because the objections and search phrasing differ by language.

A straight translation carries your English sentence structure into the other language, along with your English assumptions about what the reader is worried about. That mismatch, not grammar, is usually what costs you the enquiry.

A workable split for most Malaysian sites:

• Translate directly: product specifications, pricing structures, terms, delivery and warranty policies, technical documentation.
• Rewrite from scratch: homepage, service pages, blog articles, and anything answering “why should I trust you”.
• Rebuild rather than translate: keyword targets. BM searchers use different phrasing, so the English keyword list rarely maps one-to-one.

If you are outsourcing this, judge the writer on whether they can interview your salespeople, not on price per word. That is the same standard we apply when comparing copywriting agencies in Malaysia.

PART 5 · DEPLOY

How to Structure a Multilingual Site So Google Reads It Right

IN BRIEFGive every language its own URL, keep one language per page, and connect the versions with hreflang. Google reads the visible words on the page to work out its language, so side-by-side translations confuse it — and confused pages rank for nothing.

Google’s own guidance on managing multi-regional and multilingual sites is unusually direct on this point: use different URLs per language version, and avoid mixing languages within a page. The localized versions documentation covers how to declare the relationships.

How to set up a multilingual Malaysian website Google can read

Four decisions, in this order, cover most Malaysian small-business sites.

1. Pick one URL pattern and keep it. Subfolders such as /ms/ and /zh/ on your main domain are the simplest to maintain and keep all authority on one domain.
2. Keep one language per page. Content and navigation in the same language, no side-by-side translations, no auto-switching based on browser settings alone.
3. Add hreflang between the matching versions. Each version points to the others and to itself, so Google can serve the right one instead of guessing.
4. Localise the trust elements too. Contact hours, WhatsApp number, business registration details and payment terms belong in the language of the page, not only in the English original.

Check the paid side too. Google’s language targeting documentation explains that it infers a user’s language from account settings, query language and browsing signals. In a multilingual market, your ad language and landing-page language have to match by design, because the platform will not repair a mismatch for you. That belongs with the rest of your search visibility work.

BENCHMARK BRIEFING 4 OF 4

What Does Splitting Across Three Languages Cost You?

IN BRIEFSplitting a fixed content budget three ways does not divide results by three — it delays the compounding that makes content work at all. The model below tracks the same monthly effort spent on one language versus spread across three.

Content returns are not linear. Early pages earn little, then the same pages start supporting each other and returns accelerate. Splitting effort keeps all three tracks in the slow early phase for longer.

The split approach does eventually catch up, and it covers more of the market when it does. The question is whether your cash flow can wait the extra two to three quarters for the first meaningful return.

PART 6 · DRIVE

Which Numbers Say It’s Time to Add a Second Language?

IN BRIEFAdd the second language when the first one is producing enquiries you can already handle, and you have evidence of demand you are missing. Decide the trigger numbers now, so the decision is made by data instead of by whoever asks loudest.

Set the thresholds before you need them. Review quarterly against these four:

• First track is compounding. Organic enquiries from language one have grown for two consecutive quarters without extra spend.
• Missed-demand evidence exists. Search Console shows impressions on other-language queries, or your team is fielding enquiries in a language your site does not serve.
• Reply capacity is real. Someone can answer, quote and follow up in the new language within your normal response time — not “we will use translation tools”.
• Budget is additive, not borrowed. Funding language two by cutting language one restarts the slow phase in both.

If three of the four are true, add the language. If only one or two are, you have a first-language execution problem rather than a coverage problem. That distinction is worth raising with any shortlisted partner, including whoever you rank as the best digital marketing agency in KL.

FAQ

Frequently Asked Questions

1. Should a Malaysian business write content in English or Bahasa Malaysia first?

Settle the content language Malaysia question by looking at the language your highest-value customers already use when they enquire. That depends on whether your revenue comes from a few large business buyers, which usually points to English, or many smaller consumer jobs across the country, which usually points to Bahasa Malaysia. Check your own enquiry records before assuming.

2. Is a Chinese version of my website worth it in Malaysia?

It is worth it when referrals and enquiries already reach you in Chinese, and rarely worth it before that. The deciding factor is whether someone on your team can quote and negotiate in Chinese at your normal response speed. Without that, a Chinese page creates enquiries you cannot convert.

3. Can I use AI or Google Translate for the other languages?

Use it for the first draft, never for the published page. Machine translation handles specifications and policies acceptably, but it cannot judge which customer objection to answer first, and that judgement is what makes a service page convert. Budget for a native speaker to rewrite the argument, not just check the grammar.

4. Does putting three languages on one page hurt my SEO?

Yes, mixing languages on a single page works against you. Google determines a page’s language from its visible content, so side-by-side translations blur the signal and the page competes weakly in every language. Use one language per page, separate URLs, and hreflang to connect the versions.

5. How many pages should I translate first?

Start with the five pages that carry your enquiries, not the whole site. That usually means the homepage, your top two service pages, the pricing or process page and contact. Measure enquiry quality from those five for a quarter before translating anything else — partial coverage that converts beats full coverage that does not.

THE VERDICT

Choose One Language, Then Earn the Next

The three-language website is not the goal. It is the reward for getting one language to work, and most Malaysian businesses skip to the reward with a budget that only covers the first step.

So settle the content language Malaysia question in this order. Read your own enquiries to find the language your paying customers use. Weigh revenue concentration against search competition to pick the first track. Staff the reply capacity before you publish, keep one language per page, and set the thresholds that tell you when language two has been earned. That sequence is the whole of it, and it is the same logic running through our content strategy work.

One language done properly will out-earn three done thinly.