Rakuten Product Page Localization for Foreign Brands

A foreign brand that has run Amazon or Shopify has an instinct for what a product page is: a title, a few bullet points, a gallery, a description block, add to cart. That instinct is the single biggest reason foreign brands underperform on Rakuten Ichiba. A Rakuten product page is not the same object with Japanese text poured into it. It is built inside a different content model, ranked by a different search engine, and consumed on a different device than the one most Western pages are designed for.

This matters because the product page is where Rakuten performance is won or lost. You can pass the application, pay the fees, and open the store — our guide to selling on Rakuten as a foreign brand covers that path — and still stall out completely if the pages themselves are translated rather than localized. This article is about that second problem: what localization actually means at the page level, which parts of it are mechanical constraints you cannot negotiate, and how to decide whether to build it in-house or partner for it.

A Rakuten Product Page Is Not a Translated Amazon Listing

Start with the format expectation, because everything else follows from it. Japanese marketplace buyers expect product pages to be dense — long, image-heavy, detailed, with specification tables, usage scenarios, size and material breakdowns, and a level of visual explanation that would look overwhelming on a typical American or European storefront. On Rakuten this is not clutter; it is the trust signal. A sparse, minimalist page that would read as premium in a Western context reads as incomplete, or even suspicious, to a Rakuten shopper deciding whether an unfamiliar overseas brand is safe to buy from.

Layer onto that a set of hard field limits that shape the work before a single word is written. Rakuten's product content is split across specific fields — item name, catchcopy, PC description, smartphone description, images — each with its own character or file constraints. You are not writing "a product page." You are filling a structured template where each field does a distinct job and obeys a distinct rule. Localization means writing native Japanese copy into that structure, with the constraints understood up front, not translating an English page and hoping it fits.

The Item Name Is a Ranking Decision, Not a Label

The Rakuten item name (商品名) is the most important localization decision on the entire page, and the one most often gotten wrong by translation. Rakuten allows an item name of up to 127 full-width characters (255 half-width), and Japanese EC operator guides commonly advise using 70–100 of them. To a Western seller used to a short, clean product title, that looks like keyword stuffing. On Rakuten it is standard practice, because the item name is the primary field Rakuten's internal search reads.

Two mechanics make placement decisive. First, only roughly the first 30–50 characters of the item name reliably display in search results and category listings before truncation — so whatever has to convince a shopper to click has to live at the front. Second, Rakuten does not publish its ranking algorithm, but Japanese operator analyses consistently report that its search appears to weight keywords positioned near the start of the item name more heavily — observed practice rather than a documented rule, but a stable one. Both point to the same discipline: the item name is assembled in a deliberate order — main category keyword first, then brand, then differentiating attributes, then specifications — not written as a translated sentence.

This is why "translate our product titles" is the wrong instruction. Your English title was written for a different search engine with different mechanics. The Japanese item name has to be constructed from keyword research on how Japanese buyers actually search your category, arranged front-to-back by ranking value. A literal translation of a good English title is almost always a weak Japanese item name.

Catchcopy: 87 Characters That Do Two Jobs

Below the item name sits the catchcopy (キャッチコピー), a short promotional line capped at 87 full-width characters shared across PC, smartphone, and app. It has to do two things at once, and they pull in different directions.

Job one is search: the catchcopy is indexed by Rakuten's search engine, so keywords placed here contribute to discoverability — though with less weight than the item name and the description. Job two is conversion: it is the line that reinforces why this product, right after the item name has done its keyword work. The tension is that a catchcopy optimized purely for keywords reads like a search string, and one written purely for persuasion may waste indexed space.

There is also a display trap that punishes brands who write for the desktop view. On category listing pages, the smartphone view truncates the catchcopy to roughly 14 characters, and the app may not show it at all. So the front of those 87 characters has to carry a complete message on its own, because on the most common browsing surface that front fragment is frequently all a shopper sees. Localizing the catchcopy well means writing a line that works fully readable at 87 characters and still lands in its first 14.

The Mobile Page Is the Real Page

The largest structural mistake foreign brands make is building the PC page first and treating the smartphone page as a shrunken afterthought. The proportions are the reverse of that assumption. The dominant share of Rakuten traffic is mobile, which means the smartphone sales description is the page most of your buyers actually read, and the PC page is the secondary surface.

Rakuten gives this its own dedicated field: a smartphone-specific sales description of up to 5,120 full-width characters with a limit of 20 images (20 <img> tags). That is a real content budget, and it is separate from the PC description. Localizing only the PC layout, or auto-collapsing it into mobile, produces exactly the kind of page that looks acceptable to a manager reviewing on a laptop and reads as broken to the shopper on a phone — mistimed image breaks, text walls, tap targets in the wrong place. A properly localized Rakuten page is designed mobile-first, within the 20-image constraint, with the specification and reassurance content that Japanese buyers look for sequenced for a vertical scroll on a small screen.

Images Carry the Weight Japanese Copy Can't

On a Japanese marketplace page, images are not decoration around the copy — they frequently are the copy. Much of the specification, comparison, and reassurance content on high-performing Rakuten pages lives inside designed images with Japanese text baked in: spec tables, size charts, before/after or usage-scenario panels, material close-ups, and trust callouts. For a foreign brand, this is often the most underestimated part of localization, because it requires Japanese-language graphic design, not just writing.

The mechanics that govern this are set by R-Cabinet, Rakuten's image management system, and they impose real constraints:

• Per-image size: individual images cannot exceed 2MB; anything larger is rejected on upload.
• Images per product: up to 20 product images, matching the 20-image limit on the smartphone description.
• Storage capacity by plan: free R-Cabinet storage is 500MB on the Ganbare! and Light plans, 5GB on Standard, and effectively unlimited (5GB initial) on Mega Shop — a real planning input when your catalog carries the heavy, text-in-image assets Rakuten pages need.
• Folder limits and filenames: the base folder holds up to 5,000 images and other folders 500 each; filenames are restricted to 20 half-width characters using only letters, numbers, hyphen, and underscore — a small operational rule that trips up teams migrating an existing asset library with Japanese or long filenames.

None of these are hard on their own. Together they mean the image half of localization is a production workflow — Japanese design, export within 2MB, naming and foldering within R-Cabinet's rules — that a brand accustomed to uploading a few clean lifestyle shots is usually not staffed to run at Rakuten's expected density.

Translation Is Not Localization

Everything above assumes the thing being placed into these fields is native Japanese e-commerce copy. It is worth stating plainly why that assumption is the whole game. Machine-translated or literally-translated Japanese reads as foreign to Japanese buyers in a way that quietly depresses conversion — and, unlike a broken link or a pricing error, it rarely surfaces as an obvious problem in your analytics. The page loads, the traffic arrives, the conversion is just lower than it should be, and nothing points to why.

The gap is not primarily about grammatical correctness. Competent translation produces grammatically correct Japanese that still uses the wrong category vocabulary, misses the reassurance phrasing Japanese consumers look for, orders information in a non-native sequence, and picks keywords no Japanese shopper actually searches. Localization is written by someone who knows how the category is sold in Japan and how Rakuten buyers evaluate an unfamiliar brand — starting from the buyer and the search behavior, not from your English source text. This is the difference between a page that is in Japanese and a page that is Japanese.

Reviews Are Part of the Page

One element of the product page's performance is not something you write at all: reviews. Rakuten does not publish its ranking algorithm, but Japanese operator analyses consistently identify review count and average rating as trust signals that sit alongside sales and conversion rate among the core ranking factors, and in competitive categories — beauty, food, electronics — top-ranked products routinely carry hundreds of reviews. A newly localized page with zero reviews starts at a structural disadvantage no amount of copy quality erases on day one.

That connects back to localization in a specific way: legitimate review acquisition on Rakuten runs through order thank-you emails, insert cards, and — most durably — product and page quality that earns organic reviews. Incentivized reviews violate Rakuten's rules and are not a shortcut. So the localized page is not just the copy and images; it is the front end of a review flywheel that a good page feeds and a translated, low-conversion page starves. Our Rakuten setup review checklist covers the launch-stage details that keep this flywheel from stalling before it starts.

Build In-House or Partner? The Real Decision

The honest framing is not "can we do this ourselves" — anyone can fill the fields — but "do we have the specific combination this requires." Localizing a Rakuten page to perform, rather than merely exist, needs four capabilities at once: native Japanese e-commerce copywriting, category keyword research for Rakuten search, mobile-first Japanese image design within R-Cabinet's limits, and RMS operational fluency to assemble it correctly. Most foreign brands have none of these in-house, and a general translation vendor supplies only the first — and often only partially.

Do it in-house if…

• You have a native Japanese e-commerce specialist — not a translator — on the team who has run Rakuten pages before.
• You have Japanese-language design capacity for the text-in-image production Rakuten pages depend on.
• Your catalog is small enough that the RMS learning curve pays back over the number of pages you'll build.

Partner for it if…

• Your "Japanese capability" is a translation vendor or a bilingual staffer without Rakuten-specific experience.
• You need pages live in weeks, not after a season of learning RMS and Rakuten search conventions by trial and error.
• You want the initial build and the keyword structure done by someone who has ranked pages in your category, then to maintain routine updates internally once the template exists.

The middle path is common and usually the most economical: partner for the initial localization build and the item-name and keyword architecture, take the template in-house for ongoing updates. That front-loads the expertise where it has the highest leverage — the structural decisions that are expensive to unwind later — without paying agency rates for routine maintenance.

Common Localization Mistakes That Cost Ranking and Conversion

• Translating the item name instead of constructing it. A literal translation of a good English title wastes the highest-value ranking field on the page. Build the item name from Japanese keyword research, front-loaded, within the 127-character budget.
• Designing the PC page first. Most buyers are on the smartphone description field. A page optimized for the laptop a manager reviews it on is optimized for the minority surface.
• Treating images as decoration. On Rakuten, spec tables, size charts, and reassurance content belong in designed Japanese images — not just in a text block a Western template would carry them in.
• Writing the catchcopy for the desktop view. With category-page smartphone truncation near 14 characters, a catchcopy whose message only completes at character 60 is invisible where it matters most.
• Launching and waiting for reviews. Reviews are a ranking input, not a passive byproduct. Plan legitimate review acquisition into the launch, not after traffic disappoints.
• Assuming grammatical Japanese equals localized Japanese. Correct is not the bar. Native category vocabulary, reassurance phrasing, and Japanese information ordering are.

How LAUNOVA Helps

LAUNOVA works exclusively with foreign brands entering the Japanese e-commerce market. As a Japan ecommerce agency, we build Rakuten product pages the way the platform actually rewards — native Japanese copywriting rather than translation, item names and catchcopy structured from category keyword research, mobile-first pages within the smartphone description and 20-image limits, and text-in-image design produced to R-Cabinet's specifications. Our Rakuten Japan setup review service also audits existing pages for the localization and ranking gaps that quietly cap performance, so brands already on Rakuten can find out what a translated page is costing them before rebuilding. Pricing is quoted per brand based on catalog size, category, and scope. If your Rakuten pages are translated rather than localized — or you're about to build them and want the structure right the first time — talk to us about your Rakuten page build →

FAQ

Q: How long can a Rakuten product name be, and does length matter for search?
A Rakuten Ichiba item name allows up to 127 full-width Japanese characters (255 half-width), and Japanese EC operator guides commonly recommend using 70–100 of them. Length matters, but placement matters more: only the first roughly 30–50 characters reliably show in search and category listings, and Rakuten's search weights keywords near the front of the item name more heavily. Treat the item name as structured ranking real estate — main category keyword first, then brand, differentiators, and specifications — not as a sentence you translate from your English title.

Q: What is a catchcopy on Rakuten and how many characters do I get?
The catchcopy (キャッチコピー) is a short promotional line attached to each product, with a maximum of 87 full-width characters shared across PC, smartphone, and app. It is indexed by Rakuten search, so it contributes to discoverability, though less than the item name and description. One catch: on category listing pages the smartphone view truncates it to about 14 characters and the app may hide it entirely, so the front of the catchcopy has to carry the message on its own.

Q: Do I need separate content for the smartphone version of a Rakuten page?
In practice, yes. Rakuten provides a dedicated smartphone sales description field of up to 5,120 full-width characters with a limit of 20 images (20 <img> tags), and the overwhelming majority of Rakuten traffic is mobile. A page designed only for desktop, or one where the PC layout is machine-shrunk to mobile, reads badly on the surface most buyers actually use. Localization that ignores the smartphone field is localizing the page fewer people see.

Q: Is translating our existing Amazon or Shopify product pages enough for Rakuten?
No. Two problems compound. First, Rakuten's format expectations are different — Japanese marketplace buyers expect dense, image-heavy pages with specification tables and usage scenarios, not the sparse layouts common on Western storefronts. Second, translated copy reads as foreign to Japanese buyers in ways that depress conversion without ever surfacing as an obvious error. Rakuten localization is a rebuild in the Japanese marketplace idiom, not a language swap on an existing asset.

Q: Should we localize Rakuten pages in-house or use an agency?
It depends on whether you have native Japanese e-commerce copywriting capability and RMS operational bandwidth in-house. Translation vendors handle language but rarely the marketplace conventions, keyword strategy, and mobile-first image design that decide performance. Brands without that combination typically get faster, better-ranking pages by partnering for the initial build and the keyword structure, then maintaining routine updates internally once the template is set.