Worked sample · B2B marketing-automation platform · UK

Japan-Fit Review

10 findings · 15 min read · reviewed July 2026

A worked sample A UK-listed marketing automation vendor, early into its Japan entry. Reviewed July 2026.

This is an independent, unsolicited review of a real company's public Japanese-language site. I have anonymised the company, since the point of the sample is the method, not the target. Every figure is sourced and dated at the end. Where I could not verify something, I say so.

1. What I read

Scope. The vendor's Japanese-language site, the pages it links to, the sign-up flow, and the English original for comparison. I also looked at where the product appears outside its own site, because in Japan that is where most of the decision happens.

The reader I read as. A marketing manager at a mid-sized Japanese retailer, told to find a marketing automation tool, who has to get three or four internal parties to agree before anything is signed. This is the ordinary case in Japan, not a hard one. In a survey of Japanese buyers of business software over 3 million yen a year, around 80% said two to four departments were involved, over 80% went through two or more approval stages, and roughly 40% took six months or longer (IDEATECH, n=307, March 2026).

2. Overall

The Japanese on this site is good. It reads as written, not converted. That is not where the gap is.

The gap is that the site was built for a buyer who decides alone, on the page, in one sitting. The Japanese buyer decides somewhere else, over months, with other people in the room, using material they have copied out of the site into a document of their own. Almost everything that buyer needs for that job is either missing, or elsewhere on the site in English.

One example stands for the rest. This vendor has several Japanese customers whose adoption is a matter of public record, confirmed from the customers' own announcements. In Japanese buying research, the single strongest factor in final vendor selection is whether a case exists from a company of similar size, cited by 44.3% of respondents (IDEATECH/Demagen, n=307, March 2026). None of those Japanese customers appear on the Japanese site. The cases shown are all from the home market, and two of them are in business categories that barely exist in Japan.

The strongest asset the company already owns is not placed where the decision is made. That pattern repeats below, which is why I read this as an editing problem rather than a translation one.

3. Findings

Ten findings, grouped by layer. Each has what I observed, why it matters here with the evidence, and what I would change. They run from the structural to the surface deliberately. The most visible items are last, and they are the least important.

Layer 1 — the shelf

F1. The category the product names does not exist in this market. The product is described with a category label that, in Japanese, returns essentially only the vendor's own material. It is not a category on either of the two comparison platforms Japanese buyers actually use, and the vendor's own Japanese site uses a different label elsewhere, so the naming is not even internally consistent. Japanese B2B buyers reach comparison sites by searching a category name plus "比較" (compare); a label with no search volume and no shelf puts the product outside that path entirely. Change: lead with the category buyers actually search, in the first line of the page. A category name here is not a positioning statement, it is an address.

F2. The product is absent from where the comparison actually happens. On the main Japanese review platform the product carries a single-digit review count and a mid-table rank; on a second platform its page exists but is empty; on a third it is not listed at all; on a fourth, a third-party summary notes that its track record in Japan is shallow. This is worth separating from quality: the satisfaction score is at or above the category leaders. The problem is sample size, not sentiment. The product is not being judged poorly, it is not being judged. That platform's leading products carry review counts thirty to fifty times higher, and that is where shortlists get built. There is a second-order risk too: in a 2023 government consumer survey (n=5,544), 63.9% said they hesitate to buy even a well-rated product if negative reviews are present. With a handful of reviews, one negative is a large share of the record — and here, most of the existing reviews raise the same issue (Japanese-language coverage). Change: treat the comparison layer as a primary surface, not a by-product of the website. Register in the categories that exist, get listed where you are absent, give existing Japanese customers a route to review, and answer the standing complaints in public.

Layer 2 — the proof

F3. The vendor's Japanese proof is not on its Japanese site. The case studies and logos are carried over from the home market. Two of the featured cases are in business categories that barely exist in Japan, so there is no size or sector for a Japanese reader to map onto. Meanwhile the vendor's confirmable Japanese customers appear nowhere on the site. Two Japanese surveys point the same way: the top-cited final-selection factor is a same-size case (44.3%), with same-industry track record second (29.5%); and among 849 decision-makers at companies over 100 staff, 51% named a rich set of cases as a deciding factor, 35% specifically same-industry cases (Macromill, January 2026). Change: build a Japanese case section from the customers who have already gone public, leading with those closest in size and sector to the buyers you want. Where a customer will not be named, an anonymous case that states industry and company size still does most of the work — it is the resemblance that matters, not the brand.

Method note, kept in the public version because it is the point. Two further Japanese customer names circulate in the vendor's own Japanese material. I traced both to a single line in one of the vendor's own posts, with no confirmation from the companies, no case page, and no press coverage. I did not count them, and I would not put them in market-facing material without confirmation. Unverifiable proof is worse than no proof in a market this sensitive to it.

F4. Nothing on the site says the vendor is in Japan. (highest priority) The Japanese site's contact route leads to an English page that lists the home region and two others, but not Japan. The company page carries no Japanese location, the footer shows only the overseas entity, and the corporate social profile lists offices in several countries with no Japan entry — even though a registered Japanese entity with a Tokyo office exists. A buyer who has read the whole site in Japanese reaches the last question they need answered — is this company actually here — and the site answers no. Two reasons this matters. Practically, Japanese buyers read corporate details to get approval, not out of curiosity: in one qualitative study all twelve participants viewed the corporate site during evaluation and ten quoted from it in internal approval documents; in a survey of 221 B2B enquirers the most common reason for trusting a vendor was that the person they would deal with was shown, with information and a photograph (54.5%). Structurally, the bar for trusting an unfamiliar party is higher here — the classic Yamagishi work distinguishes general trust (an unknown counterpart is basically honest) from assurance (confidence from a known, ongoing relationship), with Japan lower on the first; the 2026 Edelman Trust Barometer places Japan last of 28 countries on its trust index. A company that shows no local address, contact, or face has closed the entrance rather than left it ajar. Change: put the Japanese entity on the Japanese site — address, contact route, the person who leads it, the customers you already have here — and route contact to a Japanese page. Smallest change on the list, and the highest-leverage, because it is the question the buyer reaches last and cannot pass.

Layer 3 — the buyer's actual job

F5. The site has one exit, and most Japanese buyers do not take it. The conversion path ends at a single form whose button says, in effect, "book now," without saying what is booked, how long it takes, who attends, or what happens next — and with no smaller step available, no downloadable price, case pack, or comparison sheet. But Japanese buyers largely do not contact vendors during evaluation: in a public-sector study of 457 SaaS users, the share who obtained information by requesting it from the vendor peaked at 20.3% across every category asked, and the report's own summary is that buyers rely on published information and terms already in hand rather than approaching the vendor. In a separate survey of 500 decision-makers, 84% reached the information that decided the purchase before speaking to a salesperson. What does open a conversation is pricing — the most common enquiry trigger, at 25.7% — and no price is published here in Japanese. Change: add the middle of the funnel. A downloadable price range, a Japanese case pack, a comparison sheet against the products this one is actually shortlisted against. Keep the demo, but let it be the third step, not the only one, and say plainly what it is.

F6. The page is built to be read, not to be cut up and circulated. The argument runs as flowing marketing copy; results appear as bare multiples with no stated conditions, baselines, or timeframes; there is no summary block, no liftable figure, no downloadable version. But the buyer's next job is to convince other people using a document of their own. Among 317 Japanese managers who had sat in internal approval meetings, 68.4% rework vendor material before sharing it — most often by pulling out figures and charts (51.2%) or taking screenshots (45.6%); 40.7% had not finished reading material they were given, the top reason being that it was too long (59.7%); and asked what a document needs to work in an approval process, 54.6% said the conclusion must be at the top, 30.9% that figures must be quotable on their own (IDEATECH/Bizibl, May 2026). A number whose conditions a buyer cannot state is a number they cannot defend when someone senior asks. Change: conclusion at the top; conditions on every claim so a figure survives being pasted into someone else's slide; stated absolutes with a baseline over bare multiples; a one-page downloadable built to be pulled apart.

Layer 4 — the channel

F7. The country's primary channel is presented as a standard feature and shipped as a preview. Japanese material lists the messaging channels in a flat sequence that includes the dominant domestic chat platform as an equal. The vendor's own help documentation describes that integration as running through an extension node, in private preview, requiring a request to a customer success manager and a chain of separate setup on the customer's side. But this is not one channel among several in Japan, it is the channel — roughly 100 million domestic monthly active users against a population near 124 million (March 2026), with paid business accounts up from 237,000 to 311,000 over two fiscal years (operator's own results). On the merchant side, native tooling for that platform already covers cart abandonment, step campaigns, ID linkage, and membership cards — precisely the automation a buyer would otherwise come to this vendor for. The exposure is not the gap itself but where it gets discovered: a buyer who commits on the channel list and finds the position different after signing hits the exact failure mode this market punishes hardest, for the structural reason in F4. Change: state the current status next to the claim, or move it out of the flat channel list until it ships as a standard channel. Being early and specific about a limitation costs a little credibility once; being found out costs the relationship.

F8. A secondary channel occupies a position it cannot earn here. SMS is presented directly after email, matching the home-market ordering. But SMS in Japan settled into authentication and notification, not promotion: the domestic SMS-delivery market is forecast smaller than the market for e-commerce site-building software, its most common business use is identity verification, its per-message cost runs several times that of the dominant chat platform and far above email, promotional sending needs prior consent under Japanese law, and SMS phishing has left real wariness (73.9% report anxiety in one survey). Change: reorder the channel list for this market — email and the dominant chat platform first, SMS positioned as what it is used for here. A change in what leads, not in the feature set.

Layer 5 — how trust is built

F9. The trust signals are ordered for a different market. The trust centre and the privacy, terms, and cookie policies reachable from the Japanese site sit in English, and the sign-up consent line, in Japanese, says the company may contact you where it has a "legitimate interest." The certification hierarchy here is close to inverted relative to Western markets: in a public-sector study of 457 SaaS users, the share treating each certification as a selection condition or reference point ran Privacy Mark 45.5%, the domestic information-disclosure certification 44.0%, ISMS 41.1%, down to SOC 2 last at 31.2% — and on the vendor side of the same study, 60.3% held ISMS and 50.4% Privacy Mark against 1.4% for SOC 2. The credential Western SaaS leads with is the one Japanese buyers weight least. Security questionnaires here are real work (dozens to hundreds of questions, two weeks to a month), and English-only material creates a second round of it. On the consent line, "legitimate interest" is a GDPR concept with no counterpart in Japan's own data-protection law; read cold in Japanese it says the company will contact you whenever it decides it has cause, and it sits directly above the submit button — the last place to introduce doubt. Change: translate the trust material and policies; state the certifications a Japanese buyer weights, in their order, and say plainly which you hold; publish a pre-filled response to the common security-questionnaire items in Japanese, which removes work from the buyer at the moment they are deciding; rewrite the consent line in terms of Japan's own framework.

Layer 6 — the words on the first screen

F10. The first screen does not say what the product is. (most visible, least load-bearing) The headline translates an English term of art into a Japanese word that carries a different, human-initiative sense, so the product category does not survive into the first line — and the category word a buyer would search is not there either (see F1). The primary button uses a word closer to a book's foreword than a first step. The subheadline reverses the English, presuming something has been lost that the original merely offered room for, and it leads with an emotional appeal ahead of specifics — where requirements-fit was the top-cited deciding factor (57%). The final call to action carries an exclamation mark that sits oddly against a purchase taking three to five months and two or more approval stages. Change: category in the first line, specific capability in the second; label buttons with the action and its consequence, not an invitation; drop the exclamation mark. Where this sits: fixing only these would change how the page reads without changing whether the product gets shortlisted, because the shortlist is assembled elsewhere. That is why they are last.

4. Two things that are not site problems

No amount of editing the site addresses these, and they bound how much the rest is worth doing.

The wedge that works at home is narrower here. Around 80% of Japanese consumer e-commerce by value runs through marketplaces rather than own-brand stores (most commonly cited third-party estimate). Marketplace sellers cannot hold the customer relationship — the largest domestic marketplace states in its own help material that a shop cannot see a customer's actual email address, and on the largest global marketplace, sellers using its fulfilment arm do not get it either. So an external marketing automation platform cannot connect to the majority of Japanese e-commerce revenue by structure, not by integration effort. The addressable market is the own-store segment, a fraction of the headline figure.

The home-market platform badge carries less here. Japan does not appear in the top ten countries by that platform's store count. In the Japanese e-commerce site-construction market, a domestic vendor has held the top share for seventeen consecutive years (43.6%, most recent figures). Enterprise retail here is not primarily that platform's conversation, so a partnership that reads as a strong credential at home reads as a niche one.

Neither is a reason to stay out of the market. Both are reasons to pick a different beachhead than the one that worked at home, and to describe the product in terms of it.

5. Priority

Ordered by effect on getting shortlisted, divided by effort.

Items 1, 2, 3, and 7 are edits to what already exists and could be done in a week. They are also four of the five highest-leverage items.

6. Sources

Government and public bodies

• Information-technology Promotion Agency, Japan. Survey on supply chain risk management for cloud services (SaaS). User survey n=457, December 2022. https://www.ipa.go.jp/security/reports/economics/scrm/t6hhco000000hg8a-att/ResearchReport.pdf
• Consumer Affairs Agency. Basic survey on consumer awareness, FY2023, n=5,544. Cited via a trade-press summary that names the survey, dates, and sample. https://netshop.impress.co.jp/node/12438
• Ministry of Economy, Trade and Industry. E-commerce market survey, 2024 results. Via secondary summary. https://www.future-shop.jp/magazine/ec-market-2024

Platform and operator primary sources

• Largest domestic marketplace, help page on its encrypted email-address service. https://ichiba.faq.rakuten.net/detail/000006811
• Dominant chat-platform operator, FY2025 results presentation (business-account figures). https://www.lycorp.co.jp/ja/ir/library/presentations/main/013/teaserItems2/04/linkList/02/link/jp2025q4_presentation.pdf
• Same operator, official-account service pages. https://www.lycbiz.com/jp/service/line-official-account/
• Store-count-by-country dataset for the home-market e-commerce platform. https://storeleads.app/reports/shopify

Buyer-behaviour surveys

• IDEATECH. Large B2B purchasing processes in Japan, n=307, March 2026. https://manamina.valuesccg.com/articles/4988
• IDEATECH and Bizibl. Internal approval documents, n=317, May 2026. https://prtimes.jp/main/html/rd/p/000000671.000045863.html
• IDEATECH and Demagen. Final vendor-selection factors, n=307, March 2026. https://prtimes.jp/main/html/rd/p/000000629.000045863.html
• Macromill. Deciding factors in external-service adoption, n=849, January 2026. https://www.macromill.com/service/report/research-report-159/
• Cone. B2B enquiry triggers, n=221, October 2025. https://cone-c-slide.com/see-sla/blog/btob-inquiry-content-decisive-factor/
• Think Partners (wib). B2B decision-makers, n=500, February 2024. https://prtimes.jp/main/html/rd/p/000000005.000064133.html
• Sairu. Qualitative study on corporate-site use in evaluation, n=12, February 2023. Directional only. https://sairu.co.jp/method/22139/

Market sizing

• ITR. Japanese SMS-delivery services market. https://www.itr.co.jp/topics/pr-20220929-1
• Fuji Chimera Research Institute. E-commerce site-construction solutions market, FY2024. https://prtimes.jp/main/html/rd/p/000000148.000007074.html
• Digital Commerce Research Institute estimate of marketplace vs own-store share, 2023. Third-party estimate, not a government figure. https://itsumo365.co.jp/blog/post-22388/

Academic

• Yamagishi, T., and Yamagishi, M. (1994). Trust and Commitment in the United States and Japan. Motivation and Emotion, 18(2), 129–166. DOI 10.1007/BF02249397
• Edelman Trust Barometer 2026, Japan report. 28 countries, 33,938 respondents, Japan n=1,202, fieldwork October–November 2025. https://www.edelman.com/sites/g/files/aatuss191/files/jp/2026-04/2026%20Edelman%20Trust%20Barometer_Japan%20Report.pdf

On confidence. Where a figure comes from a marketing-research firm's own survey rather than a public body, I have given the sample size and date so it can be weighed. Two things I deliberately left out. First, national-culture indices, because the methodology is contested in the academic literature and the argument here does not need them. Second, several widely repeated figures on message open rates in Japan, because I could not trace them to an independent source.