Thai-language SEO: word segmentation, spelling variants, URLs and bilingual sites

Thai-language SEO follows the same principles as SEO anywhere, but Thai adds problems of its own: words written without spaces, keywords spelled several ways, URLs with Thai characters, titles whose length cannot be judged by character count, and bilingual site setup. Google documents some of these clearly and has never published details on others. This guide separates the two, so decisions rest on what is actually known rather than on received wisdom.

How Google segments Thai text that has no spaces

Thai is written without spaces between words. Spaces in Thai mark phrase or sentence boundaries, not word boundaries. A search engine therefore has to segment the text into words before it can work out what a string contains.

To be plain about it: Google has not published how its systems segment Thai, which dictionary they use, or how they resolve ambiguous strings. Any article or course claiming to know "Google's Thai segmentation algorithm" is guessing. What is known are the general methods used in Thai language processing, which rely on dictionaries, statistical or machine-learning models, or a mix of these, and the problem every system faces, which is ambiguity. The textbook example is ตากลม, which can be read as ตา กลม ("round eyes") or ตาก ลม ("catch the breeze"). The system has to lean on surrounding context to decide.

That has three practical consequences:

• Write natural Thai. Do not insert spaces between words to "help" Google segment. It breaks Thai writing conventions and trips up readers, and there is no evidence it helps rankings.
• Use keywords in the form people actually type and make the surrounding context clear. An ambiguous word is easier to interpret correctly when the sentences around it are about the same subject.
• Check real results in the search results page and Search Console instead of guessing. If a page shows for the query you intended, the system understood the word, whatever method it used.

Spelling variants (ทำ/ทํา, spaced or unspaced): do they need separate pages?

No. One page that answers the topic fully should serve every spelling of the same query. Building separate pages for รับทำ seo and รับทำseo, or for ทำ and ทํา, creates pages with duplicate content that compete with each other. Several pages built to catch near-identical queries also match the doorway-page pattern described in Google's spam policies.

The origin of the ทำ/ทํา problem is more interesting than it looks. The two strings look identical but are different character sequences in Unicode. The correct ทำ uses the single character SARA AM (U+0E33). The variant ทํา is NIKHAHIT (U+0E4D) followed by SARA AA (U+0E32). On screen they look the same, but to a computer they are different text. Unicode defines NIKHAHIT plus SARA AA as the compatibility decomposition of SARA AM, so any software that applies NFKC normalization turns ทำ into ทํา. A keyword tool that does not normalize text can list the two forms as separate queries.

What works in practice:

• Write content with the standard spelling (the single SARA AM), and search your CMS for stray NIKHAHIT plus SARA AA sequences.
• If the site has internal search, normalize the two forms to one before running the query, so users still find what they are looking for.
• Search both forms on Google and compare the results pages. If the results are nearly identical, one page covers both. If they differ a lot, that is still not a reason to build duplicate pages, but you can work the form people actually use into the content naturally, for example in an FAQ.

Spacing between Thai and English words, as in รับทำ seo versus รับทำseo, follows the same logic. It is common Thai practice to put a space before and after an English word, so write it that way in content and look in Search Console at which form people type more. Google has not said it always treats the two as the same query, so check against your own site's data.

Thai or English URL slugs?

Google has not said either one ranks better. Google's URL structure documentation recommends readable words rather than long ID numbers, using the language of your audience, hyphens (-) rather than underscores (_) to separate words, and percent encoding for non-ASCII characters such as Thai in links.

The trade-offs of a Thai slug are therefore about usability rather than rankings:

• A Thai slug is instantly readable in the browser address bar and matches the language of the searcher.
• Copied into LINE, email or a document, a Thai slug often turns into a long %E0%B8%... string. Each Thai character takes 3 bytes in UTF-8, and each byte is written as 3 characters when encoded, so one Thai character becomes 9.
• Some analytics tools show the encoded form in reports, which makes them harder to read.
• An English slug is short and easy to share, but does not describe the content in Thai.

Consistency matters more than the choice. Pick one convention for the site, and do not change the URL of a page that already ranks just to switch slug language. If a change is unavoidable, set a 301 redirect from every old URL to its new one and update internal links to point straight at the new URLs.

How long can Thai titles and meta descriptions be?

Google sets no maximum character count. Its documentation says there is no limit on the length of a title element or a meta description, but both are truncated in search results as needed, typically to fit the device width. Figures like "60 characters" or "160 characters" are rough guides drawn from observation, not Google rules.

For Thai, those figures are even less reliable. Upper vowels, lower vowels and tone marks stack above or below consonants and take no extra horizontal space, yet character counters count each one like any other character. The word ที่นี่ counts as 6 characters but takes the width of 2 consonants on screen. A Thai title that counts 70 characters may display in full, while a 70-character English title may be cut off.

Practical guidance:

• Put the main keyword and the most important words at the start of the title, so they survive truncation on any screen width.
• Use a search result preview tool that measures in pixels rather than counting characters.
• Know that Google may not use your title and may generate the title link from other sources on the page, such as the h1 or anchor text, if it judges the title a poor match. Google also asks for the title to be in the same language and script as the page's main content, so a Thai page with an English or romanized title may get a different title link.
• A meta description is more likely to be used as the snippet when Google judges it describes the page more accurately than on-page text. Snippets are built mainly from page content and can change with the query, so the description is not guaranteed to appear.

How to set up a bilingual Thai-English site

Google's documentation on multilingual sites states a few clear rules:

• Use a separate URL for each language, such as /th/ for Thai, rather than switching languages with cookies or browser settings on one URL.
• Google determines a page's language from its visible content, not from the lang attribute in the code. A Thai page needs real Thai content.
• Avoid side-by-side translations on one page. Keep each page in a single language for both content and navigation.
• Do not automatically redirect users to another language based on a guess. It can stop both users and Google from reaching some versions. Offer a language switcher users click themselves.
• Do not translate only the menu and footer while leaving the main content in one language. Users end up with a page they cannot read.

Telling Google which pages are translations of each other with hreflang is a separate technical topic, and the annotations have to be correct in both directions. If English pages show up in Thai search results, or the reverse, hreflang is usually the first thing to check in an SEO audit.

Another trap on bilingual sites is translating keywords literally. The Thai queries people actually use are often not direct translations of the English keyword. For some subjects Thai users search with transliterated English, for some with pure Thai, and for some with a mix, as in the query seo ภาษาไทย itself. Thai pages need their own keyword research, not a translation of the English page's list.

Limits of Thai keyword tools

Every keyword tool gives estimates, and Thai adds limits worth knowing before you make decisions from the numbers:

• Google Keyword Planner reports average monthly searches for a keyword together with its close variants, and the figures are rounded. Slightly different spellings may be merged into one number.
• Third-party tools often list each spelling as a separate keyword, both ทำ/ทํา and spaced/unspaced forms. Each form's number then looks smaller than real demand for the topic.
• Low-volume Thai queries often do not appear in a tool's database at all, even though people do search them.
• Automatic keyword clustering depends on the tool's own segmentation, which can split Thai words wrongly and group unrelated terms together.

For an existing site, the most reliable source is the queries report in Google Search Console, because it shows the words people actually typed when your site appeared, including odd spellings no tool has. Use keyword tools to find new topics and compare rough sizes between them, then use Search Console to confirm which forms people really use.

What Google documents and what you have to test

What this means for websites in Thailand

Thai sites built on themes or platforms designed with English in mind can run into small problems that pile up: fonts that clip tone marks, Thai text that uses NIKHAHIT plus SARA AA instead of SARA AM, CMS SEO plugins warning that a Thai title is "too long" when it displays in full, or a language switcher that relies on a cookie on a single URL.

To plan SEO for the Thai market as a whole, see our SEO Thailand service, and if your business is in Bangkok and you want a team managing local rankings, see our SEO service in Bangkok.

FAQ about Thai-language SEO

Should I put spaces between Thai words so Google understands them?

No, write Thai normally without spaces between words. Google has not disclosed how it segments Thai, there is no evidence that inserting spaces helps rankings, and it certainly makes the text harder to read.

Do ทำ and ทํา need separate pages?

No, both are the same query, so keep one page written with the standard spelling (SARA AM). Duplicate pages for each spelling compete with each other and risk matching the doorway-page pattern.

How many characters should a Thai title be?

There is no fixed number, because Google does not limit title length but truncates to screen width, and Thai vowels written above or below a consonant, like tone marks, take no horizontal space. Put key words first and use a preview tool that measures in pixels.

Do Thai URLs rank better?

Google has not said Thai URLs rank better than English ones. Both work if they are readable and use hyphens between words; what matters more is consistency, and a 301 redirect whenever an existing URL changes.

A keyword tool says a Thai term has low volume. How far can I trust it?

Only for rough comparisons, because tools often split spellings into several keywords and lack data on low-volume Thai queries. Use the Search Console queries report to confirm what people actually type.

To have your Thai-language site checked against every point above, start with our SEO audit service.