How to Localize a Product Title for Coupang Search (Beyond Translation)

Why translated titles underperform on Coupang

A translated title answers the question "what is this product?" A localized title answers a different question: "what will a Korean shopper type into the search box, and does my title contain those exact strings?" Those are rarely the same thing.

Coupang's search is heavily token-driven. If a buyer searches 무선 이어폰 (wireless earbuds) and your title only says 블루투스 이어폰 (Bluetooth earbuds), you are competing for a different query — even though both phrases describe the identical product. Translation gets you a grammatically correct sentence. Localization gets you the tokens.

There is a second layer: Korean shoppers scan titles for spec and trust signals — capacity, size, certification, quantity, gift/set framing — in a fairly predictable order. A title that reads naturally in English often buries those signals past the point where the mobile app truncates.

• Translation optimizes for correctness; localization optimizes for query match plus scan-ability.
• Assume most impressions happen on mobile, where only the first portion of the title is visible before truncation.
• Your title is doing two jobs at once: retrieval (be findable) and conversion (be clickable).

Step 1: Find the token, not the translation

Almost every product category in Korea has two or three competing ways to say the same thing: a Korean-native word, an English loanword written in Hangul, and sometimes a shortened slang form. Only one of them usually carries the search volume.

Classic examples: 텀블러 (tumbler, loanword) vs. 보온병 (thermos, native) — different products to a Korean shopper, even if your English catalog calls both "insulated bottle." Or 러닝화 (running shoes) vs. 운동화 (athletic shoes), where the broader term has far more volume but far more competition. 반려동물 (companion animal, formal) vs. 강아지 (puppy/dog, colloquial) is another split where the colloquial form usually wins in search.

Do not guess. Pull the actual search demand for each variant before you write anything. This is the step most cross-border sellers skip because the data is locked behind Korean-language interfaces — it's the core of what SeoulRank surfaces in English, alongside daily rank movement for the terms you pick.

• List every plausible Korean phrasing for your product: native word, Hangul loanword, abbreviation, and category term.
• Check volume for each variant separately — never assume the dictionary word is the popular one.
• Look at how the top 10-20 listings for your target query phrase the product; consensus wording is a strong signal.
• Note whether shoppers search by product type, by use case (캠핑용 = for camping), or by problem (층간소음 = inter-floor noise).
• Keep a short list of 3-5 head/mid tokens you will commit to, and drop the rest.

Step 2: Order tokens the way Korean shoppers read them

Korean e-commerce titles follow a loose but real convention: brand → key modifier/use case → core product noun → spec → quantity/set. That's roughly the reverse of an English marketing title, which tends to lead with adjectives and benefits.

A workable skeleton: [브랜드] [용도/속성] [핵심 키워드] [사이즈/용량/색상] [수량/구성]. For example, a 500ml insulated tumbler for camping might read: BrandName 캠핑용 보온 텀블러 500ml 1개 — brand, use case, attribute, core noun, capacity, quantity.

Put your single highest-value token inside the first visible portion of the title. If the strongest query is 무선 이어폰, that phrase should appear early and intact — not split across other words.

• Front-load the one query you most want to win; treat the tail of the title as secondary coverage.
• Keep multi-word key phrases contiguous so they match the query string exactly.
• Move genuinely uninformative brand slogans and adjectives to the product detail page, not the title.
• If your brand has no recognition in Korea, don't spend early characters on it — lead with the category token.

Step 3: Get spacing, spelling, and script right

Korean spacing (띄어쓰기, word spacing) affects matching in ways English sellers don't expect. Shoppers type both 무선이어폰 and 무선 이어폰. The safest approach is to write the correctly spaced form in your main title and let the unspaced variant appear naturally elsewhere in your listing rather than jamming both into the title.

Also decide deliberately about script. Some tokens are searched in Hangul, some in Latin letters, some in both — model numbers, sizes like XL, and tech specs (USB-C, 4K) are usually typed in Latin. Capacity and dimension units follow Korean conventions: 500ml, 1.5L, 30cm, 100매 (100 sheets), 2입 or 2개 (2 pieces).

• Use the standard spaced form of key phrases; don't stack spaced and unspaced duplicates.
• Keep Latin-script tokens for model numbers, sizes, and technical standards where shoppers type them that way.
• Match Korean counter words to the category: 개 (units), 매 (sheets), 입 (pieces in a pack), 세트 (set), 팩 (pack).
• Avoid machine-translation artifacts — literal renderings of "premium quality" or "best seller" read as spam and consume characters.
• Double-check loanword romanization conventions; a mis-transliterated brand or material name is invisible to search.

Step 4: Add spec and trust tokens Korean buyers filter on

Korean shoppers frequently append qualifiers to their queries, and those qualifiers are your cheapest source of mid-tail traffic. Common ones include 대용량 (large capacity), 무선 (wireless), 접이식 (foldable), 방수 (waterproof), 국내배송 (domestic shipping), 정품 (genuine/authentic), 선물용 (for gifts), and season/use modifiers like 캠핑용, 차량용 (for vehicles), 유아용 (for infants).

Only include a qualifier if it is literally true and relevant. Regulated categories — food, cosmetics, kids' products, electrical goods — have real compliance constraints on claims, and unsupported terms invite listing suppression rather than ranking gains.

Certification and origin tokens (KC 인증, 식약처 등록) belong in the title only when they are accurate and required for the category; otherwise put them in the detail page where they still support conversion.

• Pick 2-3 truthful qualifiers that map to real searches, not five that dilute the title.
• Prefer qualifiers you can see in competitor titles across the top of the results page — that's evidence of demand.
• Skip superlatives; Korean buyers discount them and they don't help retrieval.
• Never claim certification, origin, or efficacy you can't document.

Step 5: Test, measure, and iterate on one variable at a time

Title localization is not a one-time task. Query language in Korea shifts with seasons, trends, and even TV or streaming exposure of a product category. Treat your title as a live experiment.

Change one thing at a time — swap the core noun, or move a token forward, or add a single qualifier — then watch daily rank for your target queries for one to two weeks before deciding. If you change three things simultaneously you learn nothing. Daily rank tracking matters more than weekly here, because Coupang positions move fast enough that a weekly snapshot hides the trend.

Keep a simple log: date, exact title before and after, target queries, and rank on day 0, 7, and 14. After a few cycles you'll have a category-specific playbook that is worth more than any generic advice.

• One variable per test; log the exact before/after string.
• Track rank for each target query separately — a title change can help one query and hurt another.
• Watch click-through alongside rank; a title that ranks but doesn't get clicked usually buried the spec signals.
• Re-audit token volumes seasonally, especially for gift, weather, and travel-driven categories.