Japanese Customer Support Outsourcing for Overseas Brands

Most Japan launch plans treat customer support as a line item to solve later. It is usually the first operational commitment that becomes unavoidable, because it starts the moment the first order ships and it does not wait for you to finish hiring.

The decision itself is narrow. There are three realistic ways for a foreign brand to get Japanese-language support: employ someone, buy it from a Japanese outsourcing vendor, or take it as part of an ecommerce operations retainer. What makes the choice harder than it looks is that they are not three prices for one service. They are three different services that happen to share a name.

What Japanese Buyers Actually Expect, Before You Compare Vendors

The specification has to come before the sourcing decision, and the Japanese specification is not the one most brands are running at home.

Speed is the headline attribute, not a hygiene factor. A 2023 consumer survey on ecommerce enquiry handling — 500 Japanese consumers aged 20 to 59 who had actually contacted an EC site — found response speed named the most important factor by 69.8% of respondents, ahead of accuracy of the answer (64.2%) and clarity (60.6%). Asked how quickly they wanted a reply, 14.4% said immediately, 12.2% within three minutes and 20.2% within five minutes: roughly half the sample expects an answer inside five minutes. A separate 2023 survey of people with enquiry experience found 51.0% expected a reply within one hour, and that 65.9% reduced or stopped their use of a service after a poor support experience.

Whether any brand can hit five minutes is a fair question. The operationally useful reading is different: this is a market where coverage during Japanese business hours is the product you are buying, and a once-daily batch reply from another timezone is not a cheaper version of the same thing — it is a different, worse thing.

The register is a skill, not a translation. The two hardest pieces of Japanese business writing are the refusal and the apology, and support work is largely made of both. The same 2023 survey found the leading reasons people contacted an EC site were product defects (53.9%) and returns or exchanges (49.7%) — the two categories where getting the politeness level wrong reads as contempt rather than as a language mistake.

Some of your replies are published. On Rakuten Ichiba a store's response to a review appears publicly under the store name. Support output is merchandising there, permanently attached to the product page. This is one of several places where the platform you sell on changes what support actually is — we mapped that in detail in who actually handles disputes and refunds across Japanese platforms, including which platform overrules your decision and whose balance gets debited while an argument is being settled.

There is a statutory floor under your own policy. Japan's Act on Specified Commercial Transactions gives distance-selling buyers a return right that applies where the seller has not properly published its own return terms. Whoever answers a returns enquiry needs to know what your Japanese-language policy actually says, not what your home-market policy says.

Option 1 — Hire Your Own Japanese Support Person

The strongest version of this option is real: someone whose only job is your customers, who learns your catalogue, and who is in your systems all day.

The salary. Japanese job-board aggregates put customer support roles at roughly ¥4.1–4.2 million a year on average, with an average advertised hourly rate around ¥1,262 and the largest full-time band sitting near ¥3.3–3.9 million. Those are recruitment-market aggregates rather than government statistics, and a genuinely bilingual person who can also handle coordination with an overseas head office sits above the middle of that range, not at it.

The statutory add-on. Japanese employers pay social and labour insurance on top of gross salary. For fiscal 2026 the employees' pension rate is 18.3% split equally between employer and employee (9.15% each); the national average health insurance rate at Kyokai Kenpo is 9.9%, also split; long-term care insurance adds 1.62% for employees aged 40 to 64; a new child-and-childcare support levy of 0.23% begins in April 2026; and employment insurance for a general business is 8.5 per 1,000 on the employer side against 5 per 1,000 for the worker. Add employer-only workers' accident insurance and the employer's share lands around 15% on top of gross pay — consistent with the figure we use throughout our agency versus in-house cost comparison.

So a ¥4.15 million salary is roughly ¥4.8 million in direct annual cost, about ¥400,000 a month, before a workstation, a helpdesk tool, recruiting fees, and the management time to actually direct the person.

The coverage problem is the real constraint. One employee cannot cover a calendar. Under the Labour Standards Act an employee who has worked six months with at least 80% attendance must be granted a minimum of ten days' annual paid leave, and since April 2019 the employer is obliged to ensure five of those days are actually taken. Add Japanese public holidays, sick days, and the fact that Rakuten's Super Sale and Okaimono Marathon cycles generate their enquiry spike at precisely the moment a solo hire is least replaceable, and the honest description of a one-person support function is coverage with scheduled holes in it.

Option 2 — A Japanese BPO or Call-Centre Outsourcer

Japan has a deep, mature outsourcing market for telephone and enquiry handling, and its pricing is unusually transparent by industry-guide standards. Japanese vendor guides describe three common structures:

• Phone-answering (電話代行): roughly ¥5,000–20,000 a month base plus about ¥100–200 per call, with setup around ¥10,000–40,000. This is message-taking and first-line filtering, not resolution.
• BPO-style handling: roughly ¥50,000–200,000 a month base plus about ¥400–800 per call, with setup around ¥200,000–800,000.
• Fixed monthly plans: roughly ¥100,000–500,000 a month covering an agreed call volume, with overage billed per call.

Treat these as order-of-magnitude industry references from Japanese vendor-published guides, not quotes — and note that they are built around telephone work. Marketplace message queues, email and chat are frequently scoped and priced differently, so ask explicitly which channels a quoted rate covers.

What you get is genuinely good at what it is. Native-level Japanese, trained escalation discipline, staffing that flexes with volume, and a supplier who runs this process at scale. For a brand whose enquiries are dominated by delivery status and simple order questions, this is often the correct answer and the cheapest one.

What you do not get is product and platform knowledge. A generalist agent does not have a Rakuten RMS login or Seller Central access, cannot look at your inventory sync, cannot check why a listing shows the wrong variant, and cannot answer a substantive question about ingredients, sizing, materials or Japanese labelling rules without escalating it to you. That escalation path is where the model quietly fails: the vendor answers the easy 70% at a good price and hands you back the 30% that actually needed a decision — usually the 30% attached to a refund, a defect, or a review about to be posted.

The other structural gap is that a BPO answers enquiries but does not run the store. Nobody at the call centre is updating the FAQ that would have prevented the enquiry, fixing the product page whose description caused it, or watching whether your defect rate is trending. Those tasks live in the operating rhythm described in our monthly Japan ecommerce operations checklist, and they stay with you.

Option 3 — Support Inside Your EC Operations Agency

The third route is to take support as a module of an operations retainer, from the same team already running your listings and campaigns.

The structural advantage is access. The people answering already hold the platform logins, already know why the variant is mislabelled because they built the page, and can fix the cause rather than apologise for it repeatedly. When a review needs a public reply on Rakuten, it is written by someone who understands how that reply reads next to the product copy — which is the same skill set as the page production covered in our Rakuten product page localization guide.

Japanese operation-agency pricing generally runs about ¥200,000–500,000 a month for full store operation, or a revenue share commonly in the 5–20% band, or a hybrid of the two; we break the structures down in our Japan ecommerce agency pricing models guide. Support usually sits inside that fee rather than beside it.

The limit is hours, and it is a real one. Support inside a retainer is bounded by the scope you contracted. It is not an unlimited queue, it rarely includes phone coverage, and it is not built for a sudden ten-fold spike — an agency team is sized for your normal volume, not for a viral moment or a recall. If enquiry volume is genuinely high and repetitive, a dedicated support vendor will beat an operations retainer on both price and elasticity, and any honest agency will say so.

The second limit is one we should state about ourselves: an agency that both runs your store and answers complaints about it is grading its own homework. A negative review is simultaneously a support ticket and evidence about the operator. Ask for raw enquiry logs rather than summarised counts, and ask who inside the agency sees a complaint that names the agency's own work.

The Gate Nobody Warns You About: Who May Instruct the Agent

This one catches overseas brands who assume outsourcing works the way it does at home. Japan draws a formal line between a services contract and worker dispatch. A long-standing Ministry of Health, Labour and Welfare notification — the 1986 standard distinguishing worker-dispatch business from contracted business, usually referred to by its number — sets out how to tell them apart, and the classification follows the actual working relationship, not the label on the contract. Where the client directly directs and supervises the vendor's individual workers, the arrangement can be treated as dispatch, and an unlicensed arrangement of that kind is what Japanese practice calls disguised contracting.

The practical translation for a support workflow: set the policy, the response templates, the escalation triggers, the quality standard and the reporting — put those in the specification. Do not build a workflow where your marketing manager sits in a group chat assigning individual tickets to a named agent at the vendor and correcting their wording in real time. That is not a stylistic preference; it is the distinction the classification turns on.

We are not a law firm and this is not legal advice. If you are designing an outsourced support workflow with Japanese staff, confirm the structure with a Japanese labour and social-security specialist (社会保険労務士) or a lawyer before it goes live — it is a cheap conversation compared with restructuring a live support function.

Side by Side

Do the Ticket Math Before You Pick

The comparison that settles this is cost per resolved ticket at your real volume. The arithmetic below is illustrative — it uses the sourced figures above and our own operating estimate of handling time, not a benchmark anyone has published.

Take the in-house case. About ¥400,000 a month fully loaded, against roughly 160 working hours, is about ¥2,500 an hour. A straightforward Japanese enquiry that requires checking an order in the back office and writing a properly-registered reply runs, on our own operating estimate, on the order of ten to fifteen minutes; a defect or return thread with photographs and follow-up runs considerably longer. At twelve minutes and perfect utilisation that is about ¥500 a ticket — but nobody is 100% utilised on tickets alone, and at a more realistic 50–60% that becomes roughly ¥850–1,000 per ticket.

Now put the BPO ranges beside it. At a ¥100,000 monthly base and ¥600 per unit, the outsourcer is cheaper than the fixed ¥400,000 hire until roughly 500 units a month. Most single-market foreign brands in their first two years are nowhere near that.

Which gives the honest decision rule, and it is not a cost rule:

• Under a few hundred enquiries a month, do not hire. The fixed cost of a person cannot be justified on volume, and you will be paying full price for a half-idle role while its holiday gaps still go uncovered.
• Choose by how much of your queue needs your catalogue to answer. If most enquiries are delivery status and order changes, a BPO handles them well. If most need someone who knows the product, the listing, or the platform back office, a generalist vendor will simply route them back to you at a per-unit price.
• Hire in-house when support has become a strategic surface — high volume, high repeat purchase, real product complexity, and enough of it that a full-time person is genuinely busy and a second one is on the roadmap.

Where Each Model Breaks in Practice

• Support bolted on after launch. Enquiries start with the first shipment. A store that goes live with no Japanese-language answer path accumulates unanswered messages and negative reviews during exactly the window when early ratings are being set.
• Buying a phone service for a message-queue problem. Most marketplace enquiries in Japan arrive as platform messages, not calls. A cheap phone-answering contract can look like coverage while the actual queue goes untouched.
• No agreed internal answer time. Every model degrades to the speed of your escalation reply. If the vendor or the agency has to wait two days for head office to confirm whether a return is accepted, the customer waits two days.
• Nobody owns the public reply. Review responses on Rakuten are a separate workflow from private enquiries, and they are the ones prospective buyers read. They are routinely left out of scope by both in-house hires and BPO contracts.
• Splitting support from the store. When the team answering complaints and the team fixing pages are different suppliers with no shared reporting, the same enquiry gets answered a hundred times instead of being designed out once.

Where LAUNOVA Fits, and Where It Does Not

We run Japanese customer service as part of ecommerce operations, not as a standalone contact centre. That means enquiry and review handling in native Japanese, sitting with the team that also produces the listings and runs the campaigns, so a recurring question turns into a page fix rather than a hundred replies. It is available as part of a full operations mandate through our Japan EC operations service, or as a defined module of a lighter engagement via remote Japan ecommerce operations support.

What we are not: a 24-hour call centre, an unlimited support queue, a law firm, or a tax advisor. If your volume genuinely requires round-the-clock phone coverage, a specialist Japanese BPO is the better and cheaper supplier and we will say so. Questions about employment structure or worker dispatch classification go to a Japanese labour specialist; questions about consumption tax registration are a separate matter we cover in our Japan consumption tax guide for foreign sellers. Our pricing is scoped to the work rather than published as a rate card — tell us your channels and enquiry volume and we will tell you which of the three models we think you should actually buy, including when that is not us.

The Short Version

Japanese support is a speed-and-register problem before it is a staffing problem: survey data puts response speed above accuracy in what buyers value, and on Rakuten your replies are public. Under a few hundred enquiries a month, a Japanese BPO or an agency module beats hiring on cost and on coverage. The question that actually decides between those two is how much of your queue can be answered without your catalogue and your platform logins — because everything a generalist vendor cannot answer comes straight back to you, at a per-unit price. And whichever route you take, write the requirement into the specification rather than instructing the vendor's staff directly, because in Japan that distinction has a legal edge to it.