Prices and incentive levels in this guide are taken from providers' own published pages on that date; this is a market that re-prices often, so check before you commit. This guide covers the six realistic routes, what each costs, and the specific things about Japan that break studies run by teams who have not worked there before. We sell one of the six, and that section is at the end and clearly marked. The other five are genuinely better for some studies, and we have tried to say which.
First, decide what "Japanese participants" means for your study
This sounds pedantic. It is the single most common way a Japan study produces confident, useless data. Three different populations get labelled "Japanese participants":
- People who live in Japan and speak Japanese as a first language. Almost always who you actually want: they use Japanese apps, shop in Japanese stores, and their expectations were formed by the Japanese market.
- Japanese speakers living abroad. Fluent in the language, but years removed from the market. Useful for language work, misleading for anything about behaviour.
- Non-Japanese residents of Japan. In the market, but not in the mainstream of it, and often served by a different set of products.
Groups 2 and 3 are heavily over-represented on any platform that recruits in English, for a reason that has nothing to do with the platform's honesty. If the sign-up page, the screener and the support inbox are all in English, the people who complete them are people comfortable operating in English. In Japan that is a small and unrepresentative slice. EF's 2025 English Proficiency Index put Japan 96th of 123 countries and regions, with a score of 446 — the "very low proficiency" band, and the eleventh consecutive annual decline.
Route 1 — Recruit from your own users
If you have Japanese users, this is the cheapest and fastest route and you should exhaust it before paying anyone. An in-app prompt or an email to users who match a behaviour you can see in your own data will beat any panel on relevance, because you can target on what people actually did rather than what they say they did.
The limits are real. You only reach current users, so you learn nothing about the people who never signed up or who left before you noticed. Your recruitment message and your consent flow need to be in natural Japanese — a machine-translated invitation reads as a phishing attempt and will depress your response rate more than any incentive raises it. And you still have to pay people, which means solving the payment problem below.
- Cost: the incentive, plus the engineering time for the prompt.
- Best for: understanding your own users' behaviour, churn, feature usage.
- Wrong for: market sizing, competitor users, anyone who has not heard of you.
Route 2 — Social media and communities
X (Twitter) is unusually dominant in Japan compared with most markets, and there are active communities on LINE Open Chat, note, and hobby-specific forums. For a narrow, enthusiastic audience — a particular game, a niche hobby, a specific medical condition — a well-written post in Japanese in the right community can find people no panel can reach.
It is also slow, unpredictable, and impossible to run to a schedule. Self-selection bias is severe: the people who volunteer are the people who care most, which is precisely the wrong sample if you are studying indifference or churn. And you need a native writer, because a post that reads as translated will be ignored.
- Cost: the incentive, plus a lot of your own time.
- Best for: hard-to-find enthusiast audiences, when you have weeks and no budget.
- Wrong for: anything on a deadline, anything needing a representative sample.
Route 3 — Japanese domestic panels
Japan has a large, mature research industry with panels that dwarf anything an international platform can offer in the market. Macromill reports around 1.3 million proprietary panel members and a 36 million network; INTAGE, Cross Marketing, GMO Research and Rakuten Insight all operate at comparable scale.
The catch is not quality. It is access. These companies are built for Japanese corporate clients: Japanese-language interfaces, yen invoicing, Japanese contract terms, and a sales process that assumes a purchasing department. Some have English pages and international offices — Rakuten Insight lists offices in London, Singapore, New York and elsewhere; Cross Marketing markets explicitly to companies entering Japan — but almost none publish a price.
A few published figures give the shape of the domestic market: Macromill's Interview Zero lists ¥25,000 per 30-minute interview slot. Unii Research, a self-serve Japanese platform, lists human interviews from ¥6,000 per participant and AI interviews from ¥5,000, with a five-participant minimum since its 20 August 2026 revision. Cross Marketing's Light Depth is listed at ¥175,000 for 500 samples. Rakuten Insight's AI chat interview product starts at ¥550,000 for 50 respondents. Note what those numbers have in common: they are all published on Japanese-language pages, for Japanese buyers, in yen.
- Cost: ¥5,000–¥25,000 per participant for self-serve products; quote-based for everything else.
- Best for: large samples, quantitative work, buyers who have a Japan entity or a Japanese-speaking colleague.
- Wrong for: a small team abroad that needs to move this week.
Route 4 — International recruiting platforms
Respondent, User Interviews and Prolific will all take your money today and all list Japan somewhere. This is the route most overseas teams try first, and it is worth understanding exactly what you are buying: recruitment only. You still run the interview, in a language you may not speak, and you still handle the analysis. Published list prices sit around $40–$49 per general-consumer recruit and $80–$94 for B2B, with the incentive on top.
For English-language markets, these platforms are excellent and this guide is not an argument against them. For Japan, apply the language test above, and ask the vendor directly how many Japan-resident panel members they have and in what language those people were recruited. A vendor who cannot answer that question has answered it. We have written comparisons of Respondent, User Interviews and Prolific with the current figures.
- Cost: roughly $40–$94 per recruit, plus the incentive, plus your own time to moderate and analyse.
- Best for: multi-market studies where Japan is one cell of several, teams with a Japanese-speaking moderator.
- Wrong for: a Japan-only study where you need the mainstream Japanese consumer and have nobody who speaks the language.
Route 5 — Japanese research agencies
Agencies that serve overseas clients — btrax, Wayfinder Japan, Kadence, CarterJMRN, Mitsue-Links, FieldworkHub and others — will handle everything: recruiting, screening, a bilingual moderator, simultaneous interpretation, facilities, analysis and an English report. For high-stakes generative research at the start of a market-entry decision, this is often the right answer, and no automated approach replaces a good moderator who knows the market.
Two practical notes. Every agency we checked on 23 August 2026 quotes on request; not one publishes a rate card or even a minimum engagement. And the lead time is measured in weeks, because scoping, contracting and scheduling all happen before any interview does. Domestic Japanese full-service qualitative projects are commonly described in the ¥200,000–¥1,000,000+ range per project; per participant, industry sources put recruiting alone at ¥7,000–¥15,000 before the incentive.
- Cost: quote only. Expect a five- to six-figure yen project minimum.
- Best for: high-stakes decisions, live moderation, in-person work, anything requiring a specialist B2B audience.
- Wrong for: a fast, cheap read on a specific question.
Route 6 — AI-moderated interviews
New since about 2024 and moving quickly. A language model conducts the interview, following a guide and probing based on what the participant actually says. It removes the moderator and the scheduling problem, which is what makes it cheap and what makes it available at short notice.
In Japan this is no longer novel. Unii Research has offered AI interviews since 2026. GMO Research launched an AI-interview-optimised respondent network on 1 October 2025. Macromill announced an AI interview service with Listen Labs on 4 August 2026. Cross Marketing, Rakuten Insight and others have products in the same space. Most of these are sold to Japanese buyers in Japanese, and none of the ones we checked publish a price.
One useful data point on how this lands in Japan: Outset, an AI-interview vendor, published an account of a study it fielded with GMO-Z.com Research covering 500 Japanese respondents, and reported that around 70% had never taken an AI-moderated interview before — and that participants were notably uncomfortable being recorded on camera in their own homes. That is worth knowing whichever vendor you use. It is also a reasonable argument for text-based formats in this market specifically.
The honest limits: no visual channel, so tone and hesitation are lost; no reaction to a prototype or a screen share; no improvisation outside the guide; and a text interview selects for people who like writing. It is a way to get many more conversations than you could otherwise afford, not a replacement for a skilled moderator on a hard question.
- Cost: roughly $50–$85 per completed session where prices are published; quote-based elsewhere.
- Best for: many conversations, quickly, on a well-defined question.
- Wrong for: exploratory work where you do not yet know what to ask.
What to pay a Japanese participant
Get this wrong in either direction and it costs you. Too low and you cannot fill the study; too high and you attract professional respondents. These are the industry norms as published by Japanese research providers, verified 23 August 2026:
| Format | Typical incentive |
|---|---|
| Online interview, 30 min, general consumer | ¥2,000–¥3,500 |
| Online interview, 60 min, general consumer | ¥4,000–¥8,000 |
| In-person depth interview, 60 min | ¥5,000–¥12,000 |
| Group interview | 1.2–1.5× the equivalent one-on-one rate |
| Professional or B2B respondent | roughly 1.5× the general rate, often ¥10,000–¥20,000+ |
| Phone interview, 15–30 min | ¥1,000–¥3,000 |
| Online vs in-person | online typically set 20–40% lower, since there is no travel |
Two things surprise overseas teams
Japanese incentives are substantially lower than US ones. User Interviews' own guidance, drawn from around 20,000 projects, works out at about $49.80 for a 30-minute remote consumer session — roughly $1.66 per minute. Respondent suggests around $90 per session for a broad general-population remote study. The Japanese equivalent is ¥2,000–¥3,500, which is under $25. That is a genuine cost advantage for you, and it is also a reason to be careful about how you present it: a buyer who hears "$13" thinks about data quality, not savings. Pay at the top of the local band rather than the bottom.
A single payment of ¥50,000 or more to an individual triggers Japanese withholding tax obligations (源泉徴収). This rarely bites a normal interview study, but it is the reason Japanese providers structure large or repeated payments the way they do, and it is worth knowing before you design a multi-session study with a large total.
Six things that are different about running research in Japan
1. Paying people is harder than recruiting them
Sending small yen payments to Japanese individuals from a foreign company is genuinely awkward: many Japanese consumers do not use PayPal, international transfer fees exceed the incentive itself, and gift-card systems are largely domestic. Whatever route you choose, solve this before you recruit, not after. Every Japanese provider handles it for you, and it is a real part of what they are charging for.
2. Cross-border transfer of personal data is regulated, and the rules are specific
Under Article 28 of Japan's Act on the Protection of Personal Information (APPI), transferring personal data to a party located in a foreign country generally requires prior consent, and the consent must be informed: you must disclose the destination country, an overview of that country's data-protection regime, and the safeguards the recipient has in place. Naming the country alone is explicitly insufficient. The alternatives are recipient countries recognised as having equivalent protection — currently only the EEA and the UK — or a recipient maintaining a PPC-compliant system with ongoing annual verification. An amendment bill passed the Diet on 10 July 2026 and was promulgated on 17 July 2026; it is not yet in force and must be in force no later than July 2028. Get advice; do not extrapolate from GDPR.
3. GDPR in the other direction is straightforward
For EU and UK buyers, Japan has held an adequacy decision since January 2019 (Implementing Decision (EU) 2019/419); the first periodic review concluded on 4 April 2023 and found the arrangement functioning well, and the review interval has since been extended to four years.
4. The time zone is worse than it looks
JST is UTC+9 with no daylight saving. Japanese business hours have zero overlap with US business hours in either direction — 9am–6pm in Tokyo is 5pm–2am the previous day in California. Europe fares better: Tokyo's late afternoon is Europe's morning, giving a usable window of two to three hours a day. If your study needs live sessions and your team is in North America, someone is getting up at 5am for six weeks. This is the single strongest argument for asynchronous methods on a Japan study, and it is worth deciding early rather than discovering it during scheduling.
5. Getting candid criticism takes deliberate question design
Practitioners who work in the market consistently report that Japanese participants tend to express dissatisfaction indirectly, and that a direct question — "did you find this confusing?" — reliably gets a polite yes. This is a widely-held practitioner view rather than a measured effect, so treat it as a technique adjustment, not a rule about people. What works, and works everywhere: ask about specific past behaviour rather than opinions ("tell me what you did the last time you opened the app" beats "do you like the app"); avoid questions answerable with yes; ask what someone thinks a friend would say; and give people a way to answer that is not face-to-face. Written and asynchronous formats reduce the social pressure of a live interview, which is one of the less obvious advantages of chat-based methods in this market.
6. Your interface language is a screener you did not intend to apply
Circling back to where this guide started: if any step of the participant's journey is in English, you have filtered your sample without meaning to. Check every step — the recruitment message, the screener, the consent form, the interview, the payment. One English step is enough to do it.
A checklist before you commit
- Which of the three "Japanese participant" populations do you actually need?
- In what language does the provider recruit, screen, interview and pay?
- How many Japan-resident panel members do they have, and will they tell you?
- Is the incentive at the top of the local band, or the bottom?
- Who is paying the participants, in yen, inside Japan?
- Does any personal data cross a border, and if so under which APPI Article 28 route?
- Does your method require you to be awake at 3am?
- Has a native speaker read your screener, your consent form and your questions?
Where torima.in fits
We built torima.in Global Research for one specific case: a team outside Japan needs ten conversations with mainstream Japanese consumers, this month, without hiring an agency or learning Japanese.
It is route 6 with route 3's panel. The panel is our own and it is Japanese-language — members find us, register, take part and get paid entirely in Japanese, through the domestic torima.in marketplace we have run since 2020 — so the language filter described at the top of this guide never gets applied. The ordering side is English and self-serve. You write your objective in English, approve the interview guide in English, and get English summaries, transcripts and a cross-session report. It is $49 per completed session plus a participant incentive of $14, $21, $34 that you choose — each one funding a ¥2,000–¥5,000 payment to the participant in yen — and unused sessions are refunded in full.
It is the wrong tool if you need a live moderator, a video session, someone watching a prototype, a narrow B2B audience, or a quantitative sample. Routes 3, 4 and 5 above are better at each of those.